Gut microbiota-derived nicotinamide mononucleotide alleviates acute pancreatitis by activating pancreatic SIRT3 signalling

Abstract

Background and Purpose

Gut microbiota dysbiosis induced by acute pancreatitis (AP) exacerbates pancreatic injury and systemic inflammatory responses. The alleviation of gut microbiota dysbiosis through faecal microbiota transplantation (FMT) is considered a potential strategy to reduce tissue damage and inflammation in many clinical disorders. Here, we aim to investigate the effect of gut microbiota and microbiota-derived metabolites on AP and further clarify the mechanisms associated with pancreatic damage and inflammation.

Experimental Approach

AP rat and mouse models were established by administration of caerulein or sodium taurocholate in vivo. Pancreatic acinar cells were exposed to caerulein and lipopolysaccharide in vitro to simulate AP.

Key Results

Normobiotic FMT alleviated AP-induced gut microbiota dysbiosis and ameliorated the severity of AP, including mitochondrial dysfunction, oxidative damage and inflammation. Normobiotic FMT induced higher levels of NAD⁺ (nicotinamide adenine dinucleotide)-associated metabolites, particularly nicotinamide mononucleotide (NMN). NMN administration mitigated AP-mediated mitochondrial dysfunction, oxidative damage and inflammation by increasing pancreatic NAD⁺ levels. Similarly, overexpression of the NAD⁺-dependent mitochondrial deacetylase sirtuin 3 (SIRT3) alleviated the severity of AP. Furthermore, SIRT3 deacetylated peroxiredoxin 5 (PRDX5) and enhanced PRDX5 protein expression, thereby promoting its antioxidant and anti-inflammatory activities in AP. Importantly, normobiotic FMT-mediated NMN metabolism induced SIRT3–PRDX5 pathway activation during AP.

Conclusion and Implications

Gut microbiota-derived NMN alleviates the severity of AP by activating the SIRT3–PRDX5 pathway. Normobiotic FMT could be served as a potential strategy for AP treatment.

Graphical Abstract

Abbreviations

16S rRNA

16S ribosomal RNA

AP

acute pancreatitis

CRE

caerulein

FMT

faecal microbiota transplantation

LC–MS

liquid chromatography–mass spectrometry

LY6G

lymphocyte antigen 6 complex locus G

NAD⁺

nicotinamide adenine dinucleotide

NaMN

nicotinic acid mononucleotide

NMN

nicotinamide mononucleotide

PRDX5

peroxiredoxin 5

TEM

transmission electron microscopy

TLCS

sodium taurocholate

What is already known

• Gut microbiota dysbiosis worsens the severity of AP.
• NAD⁺, a nicotinamide mononucleotide (NMN)-related metabolite, is involved in the progression of AP.

What does this study add

• Gut microbiota-derived NMN ameliorates the severity of AP.
• The NMN-activated SIRT3 attenuates the severity of AP by deacetylating peroxiredoxin 5 (PRDX5).

What is the clinical significance

• Modulation of gut microbiota or the NMN represents a strategy to limit AP.
• NMN-activated SIRT3 could be a potential target to protect against AP.

1 INTRODUCTION

Acute pancreatitis (AP) is a gastrointestinal disorder of the exocrine pancreas related to tissue damage and systemic inflammation (Pandol et al., 2007; Perez et al., 2015). Most AP cases are mild and recover spontaneously within days, but 20% of them follow a moderate to severe course with a higher mortality (Mayerle et al., 2019). The imbalance of intestinal microbiota caused by impaired pancreatic exocrine function contributes to increased intestinal barrier permeability and bacterial translocation (Frost et al., 2019; Zhu et al., 2018), which is considered a key factor in aggravating life-threatening complications (Li et al., 2020). Faecal microbiota transplantation (FMT) has been shown to mitigate tissue damage and inflammation in diverse gastrointestinal diseases, including refractory Clostridioides difficile infection (CDI) (Monaghan et al., 2021) and colitis (Burrello et al., 2018). These observations raise the possibility that FMT represents an emerging treatment to reduce the severity of AP, and further study is needed to clarify the underlying mechanisms.

FMT regulates a variety of clinical diseases by affecting metabolites. Accumulated evidence indicates that gut microbiota-derived metabolites exert profound effects on diverse gastrointestinal disorders (Lei et al., 2021; Scott et al., 2020). Among them, the short-chain fatty acids (SCFAs) and NAD⁺ (nicotinamide adenine dinucleotide)-associated metabolites ameliorate oxidative stress (Tarantini et al., 2019) and immune inflammation (Pan et al., 2019), which are the key pathogenic factors of AP. Shen et al. found that NAD⁺ participated in the pathogenesis of AP by regulating inflammation (Shen et al., 2017). Recent studies have demonstrated that gut microbiota boost host nicotinamide mononucleotide (NMN; the NAD⁺ precursor) metabolism (Shats et al., 2020). NMN supplementation has been shown to treat diabetes (Yoshino et al., 2011) and heart failure-induced inflammation and oxidative stress by promoting NAD⁺ biosynthesis in tissues (Karamanlidis et al., 2013; Yasuda et al., 2021). Thus, we hypothesized that NMN derived from the gut microbiota regulates the severity of AP. However, the mechanism of action of gut microbiota-derived NMN in regulating AP development remains unclear.

NMN exerts its effects by converting to NAD⁺, which requires NAD⁺-dependent enzymes, especially sirtuin deacetylase activity (Amano et al., 2019; Karamanlidis et al., 2013; Tarantini et al., 2019). Among the seven members of the sirtuin family, sirtuin 3 (SIRT3) is specifically localized to mitochondria and regulates mitochondrial dysfunction (Cheng et al., 2016; Gao et al., 2020; Wang et al., 2020), which may be a key causative factor in inflammatory diseases (Biczo et al., 2018; Peng et al., 2018; Xiang et al., 2021). Previous studies have shown that NMN-mediated SIRT3 activation mitigated oxidative stress (Song et al., 2018), inflammation, and adenosine triphosphate (ATP) loss in various diabetes (Caton et al., 2013) and cardiovascular diseases (Martin et al., 2017). Indeed, SIRT3 possesses robust deacetylase activity towards a series of metabolic targets, including subunits of the electron transport chain (Koentges et al., 2015), as well as enzymes involved in redox balance, and the tricarboxylic acid (TCA) cycle (Koentges et al., 2015), thereby reducing inflammation, oxidative stress and energy loss in clinical disorders (Dikalova et al., 2020). Cooley et al. (2021) found that lysine residue acetylation of nascent protein is involved in the acinar function during pancreatitis. However, the effect of SIRT3 activation induced by NMN administration on inflammation and oxidative stress in AP remains unclear.

In our study, we aim to determine the mechanism by which gut microbiota-derived metabolites ameliorate the severity of AP. Our data showed that FMT resulted in a large amount of NMN reaching the pancreas and immediate conversion to NAD⁺, in which the NAD⁺-dependent deacetylase SIRT3 deacetylated PRDX5 and increased PRDX5 levels, thereby protecting against AP-induced oxidative damage and inflammation.

2 METHODS

2.1 Animals and experimental model

All experimental procedures involving animals were carried out according to protocols approved by the Institutional Animal Ethics Committee of The First Affiliated Hospital of Harbin Medical University (Heilongjiang, China). The Institutional Animal Care and Use Committee (IACUC) number is 2020010. In addition, animal studies are reported in compliance with the ARRIVE guidelines (McGrath & Lilley, 2015; Percie du Sert et al., 2020) and with the BJP recommendations and requirements (Lilley et al., 2020). All animal experiments were conducted in accordance with the guidelines for the care and use of medical laboratory animals. Male C57BL/6 mice (7–8 weeks old weighing 20 ± 2 g) and male Wistar rats (7–8 weeks old weighing 250 ± 20 g) were obtained from the Liaoning Changsheng Biotechnology Co. Ltd. (Liaoning, China). All animals were housed in specific pathogen-free environment with controlled conditions, including a constant temperature (22 ± 2°C), 12-h light/dark cycle and 45 ± 5% humidity. The animals had free access to water and standard chow. The above animals were acclimated for 1 week before starting the experiment.

In this study, animals were assigned to different experimental groups in a blinded and random fashion according to the guidelines of the BJP. The animal groupings were blinded to the pathologists but not to the experiment operators. Caerulein (CRE)-AP was performed in mice by 12 hourly intraperitoneal injections of CRE (Sigma-Aldrich, St. Louis, MO, USA; 50 μg·kg⁻¹); the sham group received phosphate-buffered saline (PBS) (Lei et al., 2021). Mice were humanely killed by exposure to CO₂, followed by cervical dislocation, 12 h after the first injection. Sodium taurocholate (TLCS)-AP was induced in rat by retrograde injection of 3.5% TLCS solution (Sigma-Aldrich; 1 ml·kg⁻¹) into the biliopancreatic duct (Perides et al., 2010), whereas the sham group was injected with PBS. Rats were anaesthetized with 3% isoflurane and maintained with 2% isoflurane. Analgesia was achieved by administration of buprenorphine hydrochloride (1.2 mg^.kg⁻¹) 30 min before bile duct ligation surgery. The abdominal cavity was opened through a 2 cm midline incision. A microvascular clip was used to clamp the end of the hepatic duct to prevent backflow. A No.5 needle was inserted into the biliopancreatic duct through the duodenal papilla. TLCS-AP was induced in rat by retrograde injection of 3.5% sodium taurocholate solution at a volume of 0.1ml^.100g⁻¹ into the biliopancreatic duct via microinjection pump at a constant rate of 0.1 ml^.min⁻¹. After infusion, the biliopancreatic duct was clipped for 5 min at the entry site. Then, the vascular clip was removed and the incision closed with 4-0-silk suture. The sham group were injected with PBS (0.1ml^.100g⁻¹). Rats were humanely killed by exposure to CO₂, followed by cervical dislocation, 24 h after TLCS induction. Serum, pancreatic tissue, intestinal content and faeces were collected for further measurements.

2.2 Faecal microbiota transplantation

Animals were treated with sterile drinking water containing antibiotics (streptomycin [5 mg·ml⁻¹] and clindamycin [0.1 mg·ml⁻¹]) for 1 week. After the antibiotics were stopped, the animals were recolonized by oral gavage of mucus (first and second days) and faeces (third day) preparations from normal animals (Burrello et al., 2018). This protocol facilitates the implantation of the mucus-related bacteria. Mucus was scraped from intestines and diluted in saline at 1:1 ratio. Intestinal content and stool were collected and diluted in saline (50 mg·ml⁻¹). The recipient animals were orally gavaged daily for 8 consecutive weeks until they were killed by exposure to CO₂, followed by cervical dislocation, one day after the end gavage.

2.3 16S ribosomal RNA (rRNA) sequencing

The faeces and mucus scraped from the intestine were stored at −80°C until DNA extraction (MoBio, Carlsbad, CA, USA). The amplified products were separated, purified and finally sequenced on the Illumina HiSeq platform (Illumina HiSeq 2500). QIIME and a reference data set from the SILVA database were used to classify all readings to the lowest possible classification level. USEARCH V.3.10 was used to combine forward and reverse reads, and UNOISE V.3 was used to infer amplicon sequence variants (ASV). The Ribosome Database Project (RDP) classifier and SILVA 16S Ribosome Database V were used to assign the classification method.

2.4 Untargeted metabolomics study

Serum samples from the AP group and AP + FMT group were stored at −80°C. Samples were detected by liquid chromatography–mass spectrometry (LC–MS) analysis techniques. Metabolic data files generated by LC–MS were processed using Compound Discoverer to perform peak alignment, peak picking and quantification. After that, the peak value was normalized to the total spectral intensity. Then the peaks were matched with related databases to obtain accurate qualitative and relative quantitative results. The untargeted metabolomics study used a smaller n number (n < 5), which was regarded as a preliminary experiment, and no statistical analysis was performed.

2.5 Metabolite supplementation

Briefly, NMN (HY-F0004, MCE, Shanghai, China) was synthesized by the MCE company and dissolved in PBS. NMN (500 mg·kg body weight⁻¹·day⁻¹) was given intraperitoneally to animals for 28 consecutive days. The last injection was 4 h before the CRE-AP model and TLCS-AP model (Tarantini et al., 2019). Animals treated with PBS were used as the sham group. Nicotinate d-ribonucleotide (nicotinic acid mononucleotide [NaMN]) (N7764, Sigma-Aldrich) was dissolved in PBS. Mice were intraperitoneally injected with 200 μl of NaMN (33.52 mg·L⁻¹) for 28 days.

2.6 Lentiviral infection

To explore the role of SIRT3 in AP, mice were treated with lentiviral vectors harbouring full-length SIRT3 for SIRT3 overexpression, SH-SIRT3 fragments for SIRT3 knockdown or a shRNA of the negative control. On the 14, 11 and 6 days of the experiment, each mouse was intraperitoneally injected with 1 × 10⁶ infectious units (IFU). All lentiviral vectors were synthesized by Genechem (Shanghai, China).

2.7 Histopathological and immunohistochemical analysis

All procedures and analysis comply with the BJP guidelines (Alexander et al., 2018). Pancreatic samples were rapidly collected from mice and rats that were humanely killed by exposure to CO₂ followed by cervical dislocation. The mouse pancreas was obtained, fixed in 4% paraformaldehyde and further embedded in paraffin. Pancreatic sections (4 μm) were stained with haematoxylin and eosin (H&E). The pancreatic injury score was calculated as described elsewhere (Biczo et al., 2018). Immunohistochemistry of myeloperoxidase (MPO) (Abcam, Cambridge, UK; Cat #ab208670, RRID:AB_2814784), lymphocyte antigen 6 complex locus G (LY6G) (Cell Signaling Technology, USA; Cat #87048S, RRID:AB_968320) and LY6G (Bioss, China; Cat #bs-20073R, RRID:AB_881385) was performed on pancreatic slides. Pancreatic sections were incubated with the primary antibodies (1:1000) and then with biotinylated secondary antibodies (Beyotime, Shanghai, China). Pancreatic sections were processed for immunohistochemical staining by standard procedures. The ImageJ software was used to detect the number of positive cells.

2.8 Transmission electron microscopy (TEM)

Ultrastructural examination was performed by TEM. Pancreatic tissues (1 mm³) were fixed and embedded. Then, ultrathin sections were sliced into 50-μm sections on a vibrating microtome (Lecia VT1200S), dehydrated in a gradient of ethanol and further embedded in resin. The sections were stained by lead citrate and uranyl acetate. Finally, we used an electron microscope (Harbin Medical University, China) to analyse the sections.

2.9 Detection of serum amylase, lipase, LDH, TNF-α, IL-6 and NMN

Serum lipase (A054-2-1), amylase (C016-1-1) and LDH (A020-2-2) levels were examined by commercial assay kits (Nanjing Jiancheng Corp, Nanjing, China), according to the manufacturer's protocols. Serum tumour necrosis factor-α (TNF-α; CSB-E04741m) and interleukin-6 (IL-6; CSB-E04639m) concentrations were quantified using specific enzyme-linked immunosorbent assay kits (CUSABIO, Wuhan, China) (Lei et al., 2021; Pan et al., 2019). The serum NMN (MM-70919R1, MM-45732M1) level was determined by specific enzyme-linked immunosorbent assay kit (Meimian, Jiangsu, China) and detected by Varioskan LUX.

2.10 SOD, GSH, MDA, NAD⁺ and NMN assay

According to the manufacturer's protocols, GSH (G263), MDA (M496) levels and SOD (S311) activity in the pancreas were measured by a commercial analytical kit (Dojindo, Kumamoto, Japan). Pancreatic NAD⁺ (S0175) levels were tested using specific assay kits (Beyotime) and detected by Varioskan LUX. Pancreatic NMN (MM-70919R1) levels were determined by specific enzyme-linked immunosorbent assay kit (Meimian) and detected by Varioskan LUX.

2.11 Pancreatic acinar cells culture and transfection

Pancreatic acinar cells were isolated from fresh mouse pancreas as previously described (Huang et al., 2017). Pancreatic acinar AR42J cells were purchased from American Type Culture Collection (Manassas, VA, USA; Cat #CRL-1492, RRID:CVCL_0143) and grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 μg·ml⁻¹ streptomycin and 100 U·ml⁻¹ penicillin. AR42J was incubated with CCK-8+ lipopolysaccharide for in vitro (LPS; Escherichia coli serotype O111:B4, Sigma). Pancreatic acinar cells were transfected with a lentiviral vector containing full-length SIRT3 for SIRT3 overexpression, a lentiviral vector containing full-length PRDX5 for PRDX5 overexpression and respective negative controls for 5 days according to the manufacturer's protocol. AR42J cells were transfected with SIRT3 expression plasmids (Gene, Shanghai, China), PRDX5 expression plasmids, as well as the PRDX5 mutant plasmid (Genescript, Shanghai, China) or respective negative control for 48 h using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol.

2.12 Western blot analysis and immunoprecipitation

The pancreas and pancreatic acinar cells protein were extracted according to the manufacturer's instructions (Beyotime, P0133B). Western blotting was immunostained with primary antibodies against SIRT3 (Proteintech, Wuhan, China; Cat #10099-1-AP, RRID:AB_10828246), pan acetyl lysine (Abcam; Cat #ab21623, RRID:AB_446436), PRDX5 (Abcam; Cat #ab180587, RRID:AB_2904214) and β-actin (Beyotime; Cat #AA128, RRID:AB_2687938). The concentration of primary antibody was 1:1000. Next, the blots were performed with the secondary antibodies (Li-COR, Lincoln, NE, USA). Protein levels were quantified using Image Studio Lite. Total proteins were extracted in immunoprecipitation lysis buffer (Sigma-Aldrich, 87787) containing protease inhibitor cocktails (Sigma-Aldrich, 4693159001). To explore endogenous protein–protein interaction, total lysates were incubated with equal amounts of anti-PRDX5 (Abmart, Shanghai, China; Cat #T58227, RRID:AB_2904214, 1:50) and 40 μl of protein A/G agarose (Beyotime, P2108) at 4°C overnight. For exogenous immunoprecipitation assays, cell lysate containing Flag-tagged PRDX5 was incubated with anti-Flag (Beyotime, P2115) at 4°C overnight. The samples were then resuspended in 5 × sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) sample loading buffer and boiled for 10 min. The collected samples were measured by western blotting. The experimental details of western blotting were in accordance with the BJP guidelines (Alexander et al., 2018).

2.13 Immunofluorescence staining

The pancreas sections and acinar cells were fixed in 4% paraformaldehyde, then incubated with SIRT3 (Santa Cruz Biotechnology, CA, USA; Cat #sc-365175, RRID:AB_10828246) and PRDX5 (Abmart; Cat #TD12307, RRID:AB_2904214) antibodies at 4°C overnight and further incubated with secondary antibodies (Beyotime, A5021, A0423) for 1 h at room temperature. The concentration of primary antibody was 1:100. The nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (Beyotime, C1002) at room temperature for 10 min. Immunofluorescence images were examined by confocal microscope (×40, Olympus, Japan). Necrotic cell death was assessed with Hoechst (Beyotime, C1017) staining and propidium iodide (PI) (Beyotime, ST512) uptake. Acinar cells were stained with Hoechst 33258 and PI and then were measured by confocal microscopy (×40, Olympus). The total number of cells showing Hoechst staining and PI uptake was quantified for each treated group to present a percentage.

2.14 Lysine acetylation analysis

The pancreatic tissue was ground into powder with liquid nitrogen and then digested into peptides. These tryptic peptides dissolved in NETN buffer were incubated with pre-washed antibody beads. The bound peptides were eluted from the beads with 0.1% trifluoroacetic acid. Finally, the eluted fractions were combined and vacuum-dried. These peptides were cleaned with C18 ZipTips (Millipore) according to the manufacturer's instructions and then analysed by 4D LC–MS/MS.

2.15 Measurement of mitochondrial membrane potential, reactive oxygen species (ROS) and ATP

The mitochondrial membrane potential was measured by the JC-1 probe (Beyotime, C2003S). In brief, acinar cells were treated with caerulein + LPS for 6 h and then loaded with the JC-1 probe for 20 min at 37°C in the dark. Acinar cells were washed with JC-1 staining buffer, and then fluorescence intensity was examined using confocal microscope (×40, Olympus). The levels of superoxide generation were determined using the Mito-SOX probe (Invitrogen, M36008). Acinar cells were treated with caerulein + LPS for 6 h and then incubated with 5 μm of the Mito-SOX probe for 10 min at 37°C in the dark. Acinar cells were detached from cultured plates with trypsin and then examined by the BD Biosciences FACSCalibur flow cytometer. Intracellular ATP levels were examined using commercial assay kits (Dojindo, Japan, CK18) according to the manufacturer's protocols. ATP was detected by Varioskan LUX.

2.16 Flow cytometry

Fresh pancreatic tissue was minced with scissors and then digested in 1 mg·ml⁻¹ collagenase-P solution (Roche, Basel, Switzerland) for 25 min at 37°C. Subsequently, the digestion was terminated by adding washing buffer that contains fetal bovine serum. Single cell suspensions were obtained after going through 70 μm filter and stained with primary antibodies at 4°C for 60 min. For macrophages, cells were surface stained with PerCP/Cy5.5 anti-mouse CD45 (Thermo Fisher Scientific, Waltham, MA, USA; Cat #45-0451-82, 1.3:100), PE/CY7 anti-mouse F4/80 (Thermo Fisher Scientific; Cat #25-4801-82, 1:20), FITC anti-mouse CD11b (Thermo Fisher Scientific; Cat #11-0112-82, 1:50) and APC/CY7 anti-mouse CD86 (BioLegend, San Diego, CA, USA; Cat #105029, 1:40). Subsequently, these macrophages were fixed and permeabilized by using the Cell Fixation & Permeabilization Kit (Thermo Fisher Scientific, 00-5521-00) and then stained with BV650 anti-mouse CD206 (BioLegend; Cat #141723, 3:50). Finally, stained cells were analysed on Attune NxT (Thermo Fisher Scientific).

2.17 Human serum

All human investigations conformed to the principles outlined in the Declaration of Helsinki and were approved by the Ethics Committee of the First Affiliated Hospital of Harbin Medical University. The institutional review board (IRB) number is 2020XS32-02. Serum from human patients with pancreatitis (n = 15) and healthy volunteers (n = 10) were collected from the First Affiliated Hospital of Harbin Medical University. The information of all the patients is listed in Table S1.

2.18 Data and analysis

The presentation of data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and reporting (Curtis et al., 2022). No statistical data were excluded, and outliers were included in the statistical analysis. The group size selection for each protocol or study was based on the designation of group size previously published for similar experiments (Pan et al., 2019; Peng et al., 2018). By using randomization and blinded analysis, all group sizes for statistical analysis were designed to be equal. The declared group size is the number of independent values, and that statistical analysis was performed using these independent values. In addition, statistical analysis was only conducted for studies in which each group size was at least n = 5. Differences between two groups were assessed using a Student's t test (for parametric data, and Gaussian distribution) or the non-parametric Mann–Whitney U-test. Statistical analysis for multiple groups was performed by one-way analysis of variance (ANOVA) followed by Fisher's least significant difference (LSD) post hoc test (for parametric data, and Gaussian distribution), only if F achieved the level of significance P < 0.05 and there was no significant variance inhomogeneity, and the Kruskal–Wallis test was used when the data were not Gaussian distributed. Most data are shown with the original values. Bioinformatic, western blot, qPCR and immunofluorescence statistical data were normalized to control group; the Y axis was shown as fold mean of the controls. Statistical analysis was performed using GraphPad Prism 8.0 (GraphPad Prism Software, La Jolla, CA, USA). Data are expressed as the mean ± SD. Statistical significance was defined as follows: *P < 0.05, **P < 0.01 and ***P < 0.001.

2.19 Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22 (Alexander et al., 2021).

3 RESULTS

3.1 FMT alleviates AP severity by altering gut microbiota dysbiosis

To evaluate the effect of FMT on AP progression, the mucosal and faecal microbiome from normobiotic animal were transplanted into AP model by oral gavage (Figure 1a). Compared with the AP group, pancreatic histological injury was dramatically attenuated by FMT (Figures 1b and S1A). Neutrophil infiltration, as reflected by the number of myeloperoxidase-positive cells and LY6G-positive cells in the pancreas, was lower in the AP + FMT group than the AP group (Figures 1b and S9B). Consistent with these findings, serum markers of pancreatitis described by serum amylase, lipase and LDH were markedly decreased after FMT (Figures 1c and S1E–G). Ultrastructural analysis by TEM showed that FMT ameliorated mitochondrial ballooning, extensively irregular and expanded endoplasmic reticula (ERs) and nuclear fragmentation in AP model (Figure 1b). The alleviation of pancreatic oxidative damage, reflected by an increase in SOD activity and GSH level and a decrease in the MDA level, occurred in conjunction with the amelioration of mitochondrial dysfunction after FMT treatment (Figure 1d). Collectively, these findings indicate that FMT attenuates the severity of AP.

FIGURE 1

Therapeutic faecal microbiota transplantation (FMT) attenuates acute pancreatitis (AP) damage and ameliorates gut microbiota dysbiosis. (a–d) The rats were randomly divided into the following four groups: control group, AP group, AP + ABX group and AP + FMT group (n = 7 per group). ABX, antibiotics. Rats in the AP + FMT group were given mucus (first day and second day) and faeces (third day) preparations of normal rats for 8 weeks. (a) Schematic representation of FMT treatment during AP. (b) Haematoxylin and eosin (H&E) staining (scale bar = 100 μm, n = 7), immunohistochemical (IHC) staining for myeloperoxidase (MPO) (scale bar = 50 μm, n = 6) and transmission electron microscopy (TEM) (scale bar = 2 μm, n = 6), red arrows (↑): mitochondria; yellow arrows (↑): endoplasmic reticula; green arrows (↑): cell nuclei; blue arrows (↑): zymogen granules. (c) Serum amylase, serum lipase, serum LDH, n = 7. (d) SOD activity in the pancreas, GSH level in the pancreas and MDA level in the pancreas, n = 7. Data were expressed as mean ± SD. (e–i) 16S rRNA sequencing analysis in intestinal content samples from FMT-treated AP model described in Figure 1a, n = 5. (e) Alpha diversity index (based on Simpson). (f) Principal coordinate analysis (PCoA) 3D of the gut microbiome composition on the genus level. (g) Linear discriminant analysis (LDA) score identified the most differentially abundant taxa between AP and AP + FMT groups, n = 5. (h) The relative abundance of the top 10 most abundant genera in the intestinal contents of all animals. (i) The relative abundance of the genera Prevotella_9, Prevotellaceae_NK3B31_group, Prevotellaceae_UCG-001 and Coriobacteriales. n = 5. *P < 0.05, **P < 0.01, ***P < 0.001

Furthermore, 16S rRNA sequencing analysis was performed to investigate the impact of FMT on the gut microbial community in AP. FMT significantly increased the AP-induced Simpson index decline (Figure 1e). The beta diversity analysis profiled three independent clusters in principal coordinate analysis (PCoA) (Figure 1f). Firmicute and Bacteroidete constituted the most dominant phyla in gut microbiota, and FMT reduced AP-mediated increase in Firmicutes/Bacteroidetes (F/B) ratio (Figure S2A). Next, bacterial abundance analysis, linear discriminant analysis (LDA) of effect size (LEfSe) and metastats analysis were used to demonstrate overrepresentation of the Lachnospiraceae_NK4A136_group, Ruminococcus_1, Ruminococcaceae_UCG-009 and xylanophilum_groups in the AP group, whereas the Prevotella_9, Prevotellaceae_NK3B31_group, Prevotellaceae_UCG-001, Coriobacteriales, Anaerovibrio, Candidatus_Stoqueflchus and Mucispirillum presented relatively higher levels in the FMT-treated group (Figures 1g–i and S2B–D). These data suggest that FMT alleviates AP severity by altering gut microbiota dysbiosis.

3.2 Gut microbiota-induced NMN ameliorates AP severity

To explore the potential mechanism behind the beneficial effects of therapeutic FMT, the LC–MS analytical technique was used to profile serum metabolites. Data from the untargeted metabolomics study are considered as preliminary experimental results owing to small group size of n < 5. We used the untargeted metabolomics study to find that nicotinate and nicotinamide metabolism was the major enriched metabolic pathway during AP (Figure 2a). NMN, a key point in the pathway, was identified in the serum of AP model (Figure 2b). Relation correlation analysis between differential species and metabolites was performed, indicating that Prevotella_9, Prevotellaceae_NK3B31_group, Prevotellaceae_UCG-001, Coriobacteriales, Anaerovibrio, Candidatus_Stoqueflchus and Mucispirillum were positively correlated with NMN, whereas Lachnospiraceae_NK4A136_group, Ruminococcus_1, Ruminococcaceae_UCG-009 and xylanophilum_group were negatively correlated with NMN (Figure 2c). Importantly, long-term FMT treatment produced moderately more physiological serum NMN and further significantly enhanced NAD⁺ biosynthesis in the pancreas (Figure S3A–C). We speculate that long-term FMT may generate more NMN by modulating gut microbiota composition. To investigate the effects of NMN on AP, C57BL/6 mice were administered with NMN or PBS for 4 weeks (500 mg·kg⁻¹·day⁻¹) prior to being subjected to AP (Figure 2d). The histological damage in the NMN + AP group was markedly alleviated compared with the untreated group (Figure 2e). NMN-pretreated CRE-AP mice presented significant reductions in serum amylase, lipase (Figure 2f), TNF-α, IL-6 (Figure 2g) and pancreatic neutrophil infiltration (Figures 2e and S10A). Similarly, pretreatment with NMN ameliorated mitochondrial dysfunction and oxidative damage in CRE-AP mice, as measured by differences in mitochondrial ultrastructure (Figure 2e) and oxidative stress markers (Figure 2h) in pancreas. Flow cytometry analysis showed that NMN significantly decreased the number of pancreatic M1 macrophages (CD45⁺CD11b⁺F4/80⁺CD86⁺) in AP mice (Figure 2i) and had no effect on M2 macrophage (CD45⁺CD11b⁺F4/80⁺CD206⁺) infiltration (Figures 2i and S3D). NMN supplementation significantly increased the pancreatic NAD⁺ levels in the AP model (Figure S3E). In addition, rats were treated with NMN (500 mg·kg⁻¹·day⁻¹) to explore the effect of NMN on the TLCS-AP model. The histopathological injury was significantly attenuated in the presence of NMN (Figure S4A,B). NMN treatment significantly ameliorated mitochondrial ballooning and extensively irregular and expanded ERs (Figure S4A). The serum amylase and lipase were also significantly reduced in the NMN-treated group (Figure S4C,D). NaMN, a key point in the nicotinate and nicotinamide metabolism, was identified by the untargeted metabolomics study (Figure 2b). To explore the effect of NaMN on AP, NaMN (33.52 mg·L⁻¹·day⁻¹) was administered to C57BL/6 mice for 4 weeks prior to being subjected to AP. Our data revealed that the NaMN-treated group displayed no significant differences in histopathological injury, serum amylase and lipase levels during AP (Figure S4E–H). These findings indicate that the FMT-induced NMN alleviates the severity of AP by increasing pancreatic NAD⁺ levels.

FIGURE 2

Nicotinamide mononucleotide (NMN) derived from gut microbiota protects against acute pancreatitis (AP). (a, b) Untargeted metabolomics-data analysis in serum samples from the faecal microbiota transplantation (FMT)-treated AP model. (a) Bubble plot of metabolic pathway analysis for the AP group and AP + FMT group. (b) The hierarchical clustering heatmap was constructed for metabolites expression between the AP group (n = 4) and AP + FMT group (n = 3). (c) The relation correlation analysis of differentially enriched species and differential metabolites. (d–i) The mice were randomly divided into the following four groups: a control group, a control + NMN group, an AP group and an AP + NMN group. NMN (500 mg·kg⁻¹·day⁻¹) was administered by intraperitoneal injection for 4 consecutive weeks until 4 h before caerulein injection. (d) Schematic representation of NMN treatment during AP. (e) Haematoxylin and eosin (H&E) staining (scale bar = 100 μm, n = 6), immunohistochemistry (IHC) staining for myeloperoxidase (MPO) (scale bar = 50 μm, n = 6) and transmission electron microscopy (TEM) (scale bar = 2 μm, n = 6), red arrows (↑): mitochondria; yellow arrows (↑): endoplasmic reticula; green arrows (↑): cell nuclei; blue arrows (↑): zymogen granules. (f) Serum amylase, serum lipase and serum LDH, n = 6. (g) Serum tumour necrosis factor-α (TNF-α) and interleukin-6 (IL-6), n = 6. (h) SOD activity in the pancreas, GSH level in the pancreas and MDA level in the pancreas, n = 6. (i) CD86 expression in CD45⁺CD11b⁺F4/80⁺ macrophages from pancreas of caerulein (CRE)-AP mice, n = 5. *P < 0.05, **P < 0.01, ***P < 0.001

3.3 SIRT3 alleviates AP-mediated oxidative damage and inflammation in vivo and in vitro

NMN mediates its effects by converting to NAD⁺, indicating that NMN requires NAD⁺-dependent deacetylase to achieve its effects. NMN can increase the level of NAD⁺ in mitochondria, thereby reducing mitochondrial protein acetylation and improving mitochondrial function (Martin et al., 2017; Wang et al., 2020). Our results revealed that NMN administration slightly increased NMN level in AP mice (Figure S5A). As shown in Figure S5B, the mitochondrial protein extract from AP pancreas was highly acetylated, and NMN treatment significantly attenuated the AP-induced mitochondrial protein hyperacetylation (Figure S5C). Among the NAD⁺-dependent deacetylase sirtuin family, SIRT3, sirtuin 4 (SIRT4) and sirtuin 5 (SIRT5) are localized in mitochondria and have emerged as regulators of mitochondrial protein acetylation (Carrico et al., 2018). The expression of SIRT3 was significantly lower than that of SIRT4 and SIRT5 in AP (Figure S5D), suggesting that inhibition of SIRT3 activity, rather than SIRT4 and SIRT5, was responsible for the increased acetylation in the AP.

To explore the role of SIRT3 in AP, a lentiviral vector containing the full-length murine SIRT3 cDNA (OE-SIRT3) was generated. C57BL/6 mice were treated with the OE-SIRT3 or the negative control construct (OE-NC) (Figure 3a). The histopathological injury and neutrophil infiltration of pancreas were mitigated in the presence of OE-SIRT3 (Figures 3b and S10B). The serum amylase, lipase, LDH (Figure 3c) and TNF-α (Figure 3d) in the OE-SIRT3 treatment group were significantly reduced compared with the untreated group. OE-SIRT3 ameliorated mitochondrial ballooning, extensively irregular and expanded ERs and nuclear fragmentation in CRE-AP mice (Figure 3b). Consistently, OE-SIRT3 preserved SOD activity and GSH level and reduced MDA level to ameliorate AP-mediated oxidative damage of the pancreas (Figure 3e). To evaluate the role of SIRT3 in regulating AP pathogenesis in vitro, SIRT3 was overexpressed in acinar cells (Figure S6A). Overexpression of SIRT3 ameliorated the AP-induced dysfunction of mitochondrial membrane potential, as reflected by an increase in red JC-1 aggregates in cytoplasm (Figures 3f,g and S7A,B). Overexpression of SIRT3 alleviated AP-mediated acinar cell death, as demonstrated by reduced Hoechst 33258/PI staining cells (Figures 3f,g and S7A,B). Mito-SOX data showed that overexpression of SIRT3 resulted in an obvious reduction in mitochondrial ROS production (Figures 3f,g and S7A,B). Besides, SIRT3 overexpression significantly ameliorated the ATP loss during AP (Figure S7A). These findings confirm the protective effect of SIRT3 on pancreatitis toxins-induced oxidative damage, acinar cell death and inflammation in vivo and in vitro.

FIGURE 3

Sirtuin 3 (SIRT3) induced by nicotinamide mononucleotide (NMN) alleviates the severity of acute pancreatitis (AP) in vivo and in vitro. (a–e) The mice were randomly divided into the following four groups: control + OE-NC group, control + OE-SIRT3 group, AP + OE-NC group and AP + OE-SIRT3 group (n = 6 per group). SIRT3 overexpression was induced by delivery of lentivirus to mice exposed to caerulein. (a) Schematic representation of OE-SIRT3 treatment during AP. (b) Haematoxylin and eosin (H&E) staining (scale bar = 100 μm, n = 6), immunohistochemistry (IHC) staining for myeloperoxidase (MPO) (scale bar = 50 μm, n = 6) and transmission electron microscopy (TEM) (scale bar = 2 μm, n = 6), red arrows (↑): mitochondria; yellow arrows (↑): endoplasmic reticula; green arrows (↑): cell nuclei; blue arrows (↑): zymogen granules. (c) Serum amylase, serum lipase and serum LDH, n = 6. (d) Serum tumour necrosis factor-α (TNF-α), n = 6. (e) SOD activity in the pancreas, GSH level in the pancreas and MDA level in the pancreas, n = 6. (f) AR42J cells were transfected with plasmids encoding SIRT3 for 48 h and then incubated with caerulein + lipopolysaccharide for 6 h to achieve AP. The mitochondrial membrane potential was examined by JC-1, and then the red–green fluorescence ratio was calculated (scale bar = 50 μm, n = 5). The activation of the necrotic cell death pathway was determined by propidium iodide (PI) using immunofluorescence (scale bar = 100 μm, n = 5). Flow cytometric analysis of acinar cells stained with Mito-SOX dyes, the abscissa represents Mito-SOX and the ordinate represents the number of cells, n = 5. (g) Primary pancreatic acinar cells were transfected with a lentiviral vector containing full-length SIRT3 for 5 days and then incubated with caerulein + lipopolysaccharide for 6 h to achieve AP. The activation of the necrotic cell death pathway was determined by PI using immunofluorescence (scale bar = 100 μm, n = 5). The mitochondrial membrane potential was examined by JC-1, and then the red–green fluorescence ratio was calculated (scale bar = 50 μm, n = 5). The mitochondrial reactive oxygen species (ROS) was showed by Mito-SOX using immunofluorescence (scale bar = 50 μm, n = 5). *P < 0.05, **P < 0.01, ***P < 0.001

3.4 PRDX5 is deacetylated by SIRT3 and participates in pancreatic acinar cell injury

Due to the beneficial role of SIRT3 in AP progression, we tried to explore the mechanism induced by SIRT3 in AP. Pancreatic tissues derived from AP mice were subjected to microarrays analysis, and the differential gene expression profiles were identified (GSE121038). Based on mitochondrial dysfunction and oxidative damage as the key pathogenic events of AP, GO analysis was used to screen out 40 differentially expressed genes between AP and control mice (Figure S6C). SIRT3 is a NAD⁺-dependent deacetylase that responds to extensive reprogramming of mitochondrial protein acetylation, including the peroxiredoxin family (Ali et al., 2019). Venn diagrams identified 39 differentially expressed genes (Figure S6D). Previous studies demonstrated that SIRT3 deacetylates PRDX3 (peroxiredoxin 3) in intestinal epithelial cells and pancreatic β cells (Wang et al., 2020). Thus, PRDX5 (peroxiredoxin 5) might be served as a potential downstream partner of SIRT3 and may be involved in mitochondrial dysfunction and oxidative stress during AP. The SIRT3–PRDX5 interaction was confirmed by direct immunoprecipitation (Figure 4b). Immunofluorescence staining showed that SIRT3 and PRDX5 were colocalized in the mitochondria of acinar cells (Figure 4a). We explored whether the interaction affects the acetylation and protein expression of PRDX5 in acinar cells. The acetylation level of PRDX5 was markedly increased in the SH-SIRT3 group, whereas the total protein of PRDX5 was decreased (Figures 4c and 5b,c). Mass spectrometry (MS) was then performed to identify the acetylation of PRDX5. We found that four lysine residues could be efficiently acetylated, including K70, K79, K82 and K98 (Figures 4d and S6F–H). Interestingly, three of them (K79, K82 and K98) are highly conserved in different organisms (Figure S6E). These indicate that SIRT3 targets the PRDX5 for deacetylation and further enhances PRDX5 expression. To determine the effect of PRDX5 on AP in vitro, PRDX5 was overexpressed (Figure S6B). The dysfunction of the mitochondrial membrane potential, based on the JC-1 red–green fluorescence ratio, was ameliorated in PRDX5-overexpressed group (Figures 4e,f and S7C,D). Overexpression of PRDX5 alleviated AP-mediated acinar cell death, as shown by Hoechst 33258/PI staining cell reduction (Figures 4e,f and S7C,D). Mito-SOX data revealed that overexpression of PRDX5 attenuated AP-induced mitochondrial ROS production (Figures 4e,f and S7C,D). Besides, overexpression of PRDX5 increased intracellular ATP content during AP (Figure S7C). Collectively, these results indicate that SIRT3-mediated PRDX5 deacetylation increases PRDX5 expression, which impedes AP-induced mitochondrial dysfunction, oxidative damage and ATP loss.

FIGURE 4

Peroxiredoxin 5 (PRDX5) is deacetylated by sirtuin 3 (SIRT3) in acinar cells and exhibits its antioxidant and anti-inflammatory activities in acute pancreatitis (AP). (a) Colocalization of PRDX5 and SIRT3 in AR42J cells (scale bar = 12.5 μm, n = 5). (b) Co-immunoprecipitation (IP) shows the interaction between SIRT3 and PRDX5 in the pancreas of mice, n = 5. (c) The level of acetylated PRDX5, n = 5. (d) Mass spectrometry analysis of four acetylation sites on PRDX5. (e) AR42J cells were transfected with plasmids encoding PRDX5 for 48 h and then incubated with caerulein + lipopolysaccharide for 6 h to achieve AP. The mitochondrial membrane potential was examined by JC-1, and then the red–green fluorescence ratio was calculated (scale bar = 50 μm, n = 5). The activation of the necrotic cell death pathway was determined by propidium iodide (PI) using immunofluorescence (scale bar = 100 μm, n = 5). Flow cytometric analysis of acinar cells stained with Mito-SOX dye, the abscissa represents Mito-SOX and the ordinate represents the number of cells, n = 5. (f) Primary pancreatic acinar cells were transfected with a lentiviral vector harbouring full-length PRDX5 for 5 days and then incubated with caerulein + lipopolysaccharide for 6 h to achieve AP. The activation of the necrotic cell death pathway was determined by PI using immunofluorescence (scale bar = 100 μm, n = 5). The mitochondrial membrane potential was examined by JC-1, and then the red–green fluorescence ratio was calculated (scale bar = 50 μm, n = 5). The mitochondrial reactive oxygen species (ROS) was showed by Mito-SOX using immunofluorescence (scale bar = 50 μm, n = 5). *P < 0.05, **P < 0.01, ***P < 0.001

FIGURE 5

Therapeutic faecal microbiota transplantation (FMT)-mediated up-regulation of nicotinamide mononucleotide (NMN) activates the sirtuin 3 (SIRT3)–peroxiredoxin 5 (PRDX5) pathway. (a) The mice were randomly divided into the following four groups: control + OE-NC group, control + OE-SIRT3 group, AP + OE-NC group and AP + OE-SIRT3 group. SIRT3 overexpression is induced by delivery of lentivirus to acute pancreatitis (AP) mice. Western blotting for SIRT3 and PRDX5 in pancreas, n = 5. (b) The mice were randomly divided into the following four groups: a control + SH-NC group, a control + SH-SIRT3-1 group, an AP + SH-NC group and an AP + SH-SIRT3-1 group. SIRT3 silencing was performed with intraperitoneal injection of lentiviral vector in C57BL/6 mice. Western blotting for SIRT3 and PRDX5 in pancreas, n = 5. (c) The mice were randomly divided into the following four groups:control + SH-NC group, control + SH-SIRT3-2 group, AP + SH-NC group and AP + SH-SIRT3-2 group. SIRT3 silencing was induced by delivery of SH-SIRT3-2 to AP mice. Western blotting for SIRT3 and PRDX5 in pancreas, n = 5. (d) Immunofluorescence staining of pancreas was performed to measure the localization and expression of SIRT3 (red) and PRDX5 (green) after FMT treatment, scale bar = 50 μm, n = 5. (e) Western blotting of SIRT3 and PRDX5 after FMT treatment, n = 5. (f) The expression of acetylated PRDX5 after FMT treatment, n = 5. (g) The level of acetylated PRDX5 after NMN treatment, n = 5. (h) Immunofluorescence staining of pancreas was used to detect the localization and expression of SIRT3 (red) and PRDX5 (green) after NMN treatment, scale bar = 50 μm, n = 5. (i) Western blotting of SIRT3 and PRDX5 after NMN treatment, n = 5. *P < 0.05, **P < 0.01, ***P < 0.001

3.5 FMT-induced NMN activates the SIRT3–PRDX5 pathway in AP

To validate the role of the SIRT3–PRDX5 pathway in the AP model, lentiviral vectors were delivered to the AP model. AP-induced SIRT3 reduction was reversed by SIRT3 overexpression (Figures 5a and S8A). Western blot and qPCR results showed that PRDX5 expression was significantly increased in the OE-SIRT3-treated AP group (Figures 5a and S8A). In addition, SH-SIRT3 lentiviral vector significantly decreased SIRT3 and PRDX5 expressions in the AP model (Figures 5b,c and S8B,C). The SIRT3–PRDX5 pathway may be involved in AP progression. To explore the effect of gut microbiota-induced NMN on regulating the SIRT3–PRDX5 pathway, the expressions of SIRT3 and PRDX5 were tested. Immunofluorescence and western blot showed that SIRT3 and PRDX5 were increased by FMT treatment (Figure 5d,e). Correspondingly, the acetylation level of PRDX5 was significantly reduced after FMT treatment (Figure 5f). Consistently, NMN treatment significantly increased SIRT3 and PRDX5 levels (Figure 5h,i). NMN dramatically decreased PRDX5 acetylation in AP mice (Figure 5g). To confirm that gut microbiota-induced NMN specifically exerts beneficial effects through SIRT3, the expressions of SIRT1 and SIRT2 were examined. Our data showed that SIRT1 was decreased in AP and that FMT slightly increased the expression of SIRT1 without statistical significance. In addition, AP induced a tendency of increased SIRT2 expression without statistical significance. FMT had no effect on SIRT2 expression (Figure S9A). Taken together, these indicate that gut microbiota-induced NMN activates the SIRT3–PRDX5 pathway in AP model.

3.6 NMN protects against AP injury in a SIRT3-dependent manner

To clarify whether NMN exerts the protective effect in a SIRT3-dependent manner, the effect of NMN on SIRT3-knockdown mice was examined. Knockdown of SIRT3 aggravated AP injury and abrogated the beneficial effect of NMN, as demonstrated by histopathological and myeloperoxidase-positive cells (Figure 6a), LY6G-positive cells (Figure S10C), serum amylase, lipase, LDH (Figure 6b), TNF-α and IL-6 (Figure 6c). The mitochondrial dysfunction and oxidative damage of pancreas, based on the pancreatic ultrastructure (Figure 6a), the activity of SOD and the level of GSH and MDA (Figure 6d), were exacerbated after treatment with SH-SIRT3, whereas no difference was found between the NMN and control groups in SIRT3-decreased mice. These findings suggest that the protective effect of NMN partially depends on the presence of SIRT3 activity.

FIGURE 6

Nicotinamide mononucleotide (NMN) prevents acute pancreatitis (AP) damage in a sirtuin 3 (SIRT3)-dependent manner. (a–d) The mice were randomly divided into the following six groups: control group, AP + SH-NC group, AP + SH-SIRT3-1 group, AP + SH-SIRT3-1 + NMN group, AP + SH-SIRT3-2 group and AP + SH-SIRT3-2 + NMN group (n = 6 per group). SIRT3 silencing was performed with intraperitoneal injection of lentiviral vector in C57BL/6 mice. (a) Haematoxylin and eosin (H&E) staining (scale bar = 100 μm, n = 6), immunohistochemistry (IHC) staining for myeloperoxidase (MPO) (scale bar = 50 μm, n = 6) and transmission electron microscopy (TEM) (scale bar = 2 μm, n = 6), red arrows (↑): mitochondria; yellow arrows (↑): endoplasmic reticula; green arrows (↑): cell nuclei; blue arrows (↑): zymogen granules. (b) Serum amylase, serum lipase and serum LDH, n = 6. (c) Serum tumour necrosis factor-α (TNF-α) and interleukin-6 (IL-6), n = 6. (d) SOD activity in the pancreas, GSH level in the pancreas and MDA level in the pancreas, n = 6. *P < 0.05; **P < 0.01; ***P < 0.001 and not significant (NS) P > 0.05

3.7 The level of NMN and AP markers in the serum of clinical patients

To investigate the clinical relevance of NMN during AP, NMN levels and clinical markers, including serum amylase, white blood cell (WBC), neutrophil and C-reactive protein (CRP) levels, were measured in healthy volunteers and AP patients. Serum NMN was significantly reduced, and NMN levels were negatively correlated with WBCs and neutrophils in serum (Figure 7a–d). These suggest that NMN may be involved in predicting the severity of AP patients (Figure 8).

FIGURE 7

The serum nicotinamide mononucleotide (NMN) is closely correlated with acute pancreatitis (AP) markers in clinical samples. (a) The heatmap showed the concentration of NMN and AP clinical markers in normal people (n = 10) and patients with AP (n = 15). (b) The heatmap showed the relationship between NMN expression and AP markers in clinical samples. Correlation analysis of serum NMN with white blood cells (WBC) (c) and neutrophils (d). Pearson correlation test was used to evaluate the correlation. CRP, C-reactive protein; PCT, procalcitonin; SAP, severe acute pancreatitis

FIGURE 8

Gut microbiota-derived nicotinamide mononucleotide (NMN) alleviates acute pancreatitis (AP) by activating the sirtuin 3 (SIRT3)–peroxiredoxin 5 (PRDX5) pathway. Faecal microbiota transplantation (FMT) alleviates AP-induced gut microbiota dysbiosis and elevates the concentration of gut microbiota-derived NMN. Large amounts of NMN reach the pancreas and are further converted to NAD⁺ (nicotinamide adenine dinucleotide), in which the NAD⁺-dependent deacetylase SIRT3 deacetylates PRDX5 and increases PRDX5 levels, thereby reducing inflammation and oxidative damage during AP. CRE/TLCS, caerulein/sodium taurocholate

4 DISCUSSION

Our study demonstrates that gut microbiota-derived metabolites alleviate oxidative damage and inflammation in AP. We made the following observations: (1) FMT ameliorates gut microbiota dysbiosis and is considered a promising approach for AP treatment. (2) Gut microbiota-derived NMN protects from AP-mediated oxidative damage and inflammation by increasing pancreatic NAD⁺ levels. (3) NAD⁺-dependent mitochondrial deacetylase SIRT3 overexpression attenuates AP severity, whereas SIRT3 deficiency exacerbates AP and abrogates the beneficial effects of NMN. (4) SIRT3 deacetylates PRDX5 and enhances PRDX5 expression, which ameliorates AP-induced oxidative damage and pancreatic acinar cell necrosis. (5) FMT-induced NMN contributes to SIRT3–PRDX5 pathway activation. Gut microbiota-derived NMN transports to pancreas and converses to NAD⁺, in which NAD⁺-dependent deacetylase SIRT3 deacetylates PRDX5 and increases its levels. As a result, AP-induced oxidative damage and inflammation are ameliorated.

AP results in gut microbiota dysbiosis (Lei et al., 2021; van den Berg et al., 2021). The imbalance of gut microbiota increases gut permeability and bacterial translocation, which results in the life-threatening complications, including systemic inflammatory response syndrome (SIRS) and multiple organ dysfunction syndrome (MODS) (Armacki et al., 2018; Souffriau et al., 2020). Recent studies show that the phenotype of gastrointestinal diseases is aggravated after treatment with dysbiotic FMT (mucosal and faecal samples from gastrointestinal diseases), including colitis and AP (Lei et al., 2021; Tang et al., 2021). In contrast, the amelioration of gut microbiota dysbiosis induced by normobiotic FMT (mucosal and faecal samples from healthy volunteers) has been recognized as an effective treatment of many clinical diseases, such as refractory CDI and colitis (Burrello et al., 2018; Geng et al., 2018). However, far less is known about the efficacy of normobiotic FMT for AP progression. Our study found that normobiotic FMT significantly attenuated AP-induced microbiota dysbiosis and ameliorated the severity of AP.

Metabolites derived from the gut microbiota exert protective effects in a variety of gastrointestinal diseases (Nagao-Kitamoto et al., 2020; Singh et al., 2019). Recent studies reveal that the gut microbiota boosts mammalian host NAD⁺ metabolism (Blacher et al., 2019; Shats et al., 2020). Consistently, our study found that FMT increased serum NMN levels and pancreatic NAD⁺ expression. Gut microbiota–NMN interaction leads to reciprocal regulation. NMN has been demonstrated to enhance the structure of the gut microbiota and intestinal barrier integrity (Huang et al., 2021). Furthermore, NMN administration increases NAD⁺ biosynthesis in the pancreas, liver and white adipose tissue (WAT), thereby reducing oxidative stress and inflammatory responses and enhancing hepatic insulin sensitivity in diet and age-induced diabetic mice (Yoshino et al., 2011). Shen et al. have reported that NAD⁺ is significantly down-regulated in AP (Shen et al., 2017). Here, we observed that NMN administration increases serum NMN levels and pancreatic NAD⁺ levels, implying that NMN might be converted into NAD⁺ in the pancreas of the AP model. Importantly, pretreatment with NMN simulated the protective effect of FMT on AP injury. These indicate that gut microbiota-derived NMN alleviates AP-mediated mitochondrial dysfunction, oxidative damage and inflammation by enhancing pancreatic NAD⁺ biosynthesis. However, short-term treatment of NMN failed to ameliorate the pathological damage of CRE-AP in mice (He et al., 2021). We speculate that different roles of NMN in AP may depend on the time-line of treatment and the concentration of NMN-converted NAD⁺ in the pancreas.

NMN exerts its beneficial effects by conversion to NAD⁺, and then its protective effects will require NAD⁺-mediated sirtuin deacetylase (Karamanlidis et al., 2013; Martin et al., 2017; Yasuda et al., 2021). Mitochondrial dysfunction, a key pathogenic event of AP, is partly regulated by post-translational modifications of mitochondrial proteins, including acetylation, phosphorylation, O-GlcNAcylation and succinylation (Biczo et al., 2018). In this study, we observed that mitochondrial protein extracted from the AP pancreas was hyperacetylated. NMN treatment significantly reduced the mitochondrial protein acetylation by promoting pancreatic NAD⁺ deacetylase activity, thereby attenuating mitochondrial dysfunction, oxidative damage and inflammation in AP mice. Among the NAD⁺-dependent sirtuin deacetylase family, SIRT3 is specifically localized in mitochondria and serves as the main regulator of mitochondrial protein acetylation (Brown et al., 2014). We found that the expression of SIRT3 was lower than SIRT4 and SIRT5. The beneficial effects of SIRT3 on oxidative stress and energy metabolism have been extensively verified (Carrico et al., 2018; Dikalova et al., 2020). In our study, overexpression of SIRT3 alleviated AP by ameliorating inflammation and oxidative damage, as well as rescuing mitochondrial dysfunction in vivo and in vitro. SIRT3 may be a potential therapeutic target for AP treatment. Accumulating evidence indicates that NMN increases the level of SIRT3 and reduces mitochondrial oxidative damage and inflammation in heart diseases (Karamanlidis et al., 2013; Martin et al., 2017). Indeed, our results suggest that NMN significantly enhances the activity of SIRT3, thereby reducing the severity of AP. Correspondingly, depletion of SIRT3 was more sensitive to pancreatitis toxins and abrogated the protective role of NMN in AP. These indicate that NMN attenuates AP severity in a SIRT3-dependent manner.

Lysine acetylation acts as a major post-translational modification that contributes to several pathogenesis (Rardin et al., 2013). Guy E. Groblewski et al. found that deficient ER acetyl-CoA imported into pancreatic acinar cells leads to pancreatitis (Cooley et al., 2021). Our study indicates that mitochondrial protein acetylation was elevated in AP. Based on the beneficial effect of SIRT3 (the main regulator of mitochondrial protein acetylation) on AP, we further explored the downstream acetylation target of SIRT3. Large-scale proteomic methods (Ali et al., 2019) and GEO DataSets were used to identify PRDX5 as a potential downstream partner of SIRT3, which was confirmed by western blot. Previous acetyl proteomics identified PRDX5 as a potential mitochondrial acetylated protein in SIRT3 knockout mice (Ali et al., 2019; Rardin et al., 2013), which was also confirmed by MS. Protein lysine acetylation may regulate protein–protein interactions and further affect various protein properties, including protein activity and expression (Kim et al., 2006). Our study found that SIRT3 directly deacetylated PRDX5 and promoted the up-regulation of PRDX5 protein. Previous studies demonstrated the effectiveness of PRDX5 in preventing oxidative stress, calcium overload and inflammation. Consistently, our results demonstrated that overexpression of PRDX5 alleviated the toxicity of pancreatitis by ameliorating oxidative damage, the inflammatory response and mitochondrial dysfunction of acinar cells.

NMN, the NAD⁺ precursor, is regarded as a supplement to promote energy metabolism, regulate glucose and rescue the metabolic complications of aging (Mills et al., 2016). NMN also enhances muscle insulin sensitivity in overweight or obese pre-diabetic women (Yoshino et al., 2021). The commensal microbe-derived NAD⁺ metabolism is involved in amyotrophic lateral sclerosis (ALS) and acute kidney injury (AKI) (Brown et al., 2014; Zhu et al., 2021). To investigate the clinical significance of NMN in AP patients, the correlation between serum NMN and serum inflammatory markers was calculated. AP was manifested by decreased NMN levels and increased serum inflammatory markers, which were negatively correlated, suggesting its potential role as a non-invasive marker for evaluating AP severity.

In summary, our study reveals that FMT-mediated alleviation of gut microbiota dysbiosis ameliorates AP severity and further illustrates the beneficial effects of gut microbiota-derived NMN on AP. Mechanistically, NMN-activated SIRT3 deacetylates PRDX5 and increases its expression, thereby reducing oxidative damage and inflammation. These provide a new strategy for AP treatment.

ACKNOWLEDGEMENTS

This study was supported by the National Natural Science Foundation of China (Nos 81871974, 82070658, 81800572, 82270666), the Natural Science Foundation of Heilongjiang Province (TD2021H001) and the Youth Innovation Talent Training Program of the General Undergraduate Colleges and Universities in Heilongjiang Province (UNPYSCT-2020157).

AUTHOR CONTRIBUTIONS

Li-Wei Liu: Conceptualization (equal); data curation (equal); formal analysis (equal); methodology (equal); writing-original draft (lead); writing-review and editing (equal). Yu Xie: Conceptualization (supporting); data curation (equal); formal analysis (equal); funding acquisition (supporting); investigation (supporting); methodology (supporting); writing-original draft (equal). Guan-Qun Li: Conceptualization (supporting); data curation (equal); formal analysis (supporting); methodology (equal); resources (equal); validation (equal); writing-original draft (equal); writing-review and editing (supporting). Tao Zhang: Data curation (equal); formal analysis (supporting); methodology (supporting). Yu-Hang Sui: Data curation (equal); formal analysis (supporting); investigation (supporting); methodology (supporting); resources (supporting). Zhong-Jie Zhao: Investigation (equal); methodology (supporting); resources (supporting). Yang-Yang zhang: Investigation (equal); methodology (supporting); resources (supporting). Wen-Bo Yang: Data curation (supporting); formal analysis (equal). Xing-Long Geng: Data curation (supporting); formal analysis (equal); resources (supporting). Dong-Bo Xue: Data curation (supporting); formal analysis (equal); methodology (supporting). Hua Chen: Conceptualization (supporting); formal analysis (equal); methodology (supporting). Yong-Wei Wang: Data curation (supporting); formal analysis (equal). Tianqi Lu: Investigation (equal); methodology (supporting); resources (supporting). Liren Shang: Investigation (equal); resources (supporting). Zhibo Li: Investigation (equal); methodology (supporting); software (supporting). Le Li: Conceptualization (lead); data curation (equal); funding acquisition (supporting); writing-original draft (supporting); writing-review and editing (equal). Bei Sun: Conceptualization (equal); funding acquisition (lead); project administration (lead); supervision (lead); writing-original draft (supporting); writing-review and editing (lead).

CONFLICT OF INTEREST

The authors have declared that no competing interest exists.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design and Analysis, Immunoblotting and Immunochemistry, and Animal Experimentation, and as recommended by funding agencies, publishers and other organizations engaged with supporting research.

Open Research

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding authors upon reasonable request. Some data may not be made available due to privacy or ethical restrictions.