Isoflurane was obtained from RWD Life Science. Anhydrous ethanol analytical reagent (AR), xylene (AR), paraffin (56–58°C), water-soluble eosin Y staining solution, Tween 20, sodium chloride (AR), trichloromethane (AR) and 3% H₂O₂ were purchased from Sinopharm Chemical Reagent Co., Ltd. Hematoxylin was obtained from MilliporeSigma. Dimethyl sulfoxide (DMSO), 10 mM of phosphate-buffered saline buffer (PBS), loading buffer, 4′,6-diamidino-2-phenylindole (DAPI) solution, anti-fluorescence attenuation sealant, IL-4, IL-6, IL-10, IL-1β and IL-1α, enzyme-linked immunosorbent assay (ELISA) kits and H₂O₂, malondialdehyde, Fe²⁺ and mitochondrial respiratory chain complex activity assay kits were purchased from Beijing Solarbio Science & Technology Co., Ltd. GSH/GSSG detection kit was from Nanjing Jiancheng Bioengineering Institute. Protease phosphatase inhibitor mixture, RIPA lysate (strong), BCA protein assay kit, JC-1 mitochondrial membrane potential assay kit and dihydroethidium (DHE) assay kit were purchased from Shanghai Biyuntian Biotechnology Co., Ltd. The details of the antibodies used are listed in Table SI.

Male C57 BL/6 mice aged 6–8 weeks, weighing 22–25 g, free from specific pathogens (n=90), were purchased from Jiangsu Huachuang Xinnuo Pharmaceutical Technology Co., Ltd. The mice were housed at 23°C and 50% relative humidity on a 12/12-h light/dark cycle. Basic feed was processed by Beijing Huanyu Zhongke Biotechnology according to the national standard (GB 14924.3-2010). The model of sepsis was induced by performing cecal ligation and perforation. Briefly, the mice were anesthetized using 4% isoflurane, followed by a sterile midline laparotomy of ~2 cm to expose the cecum. Half of the distal end of the cecum was ligated at its center, and then a 21-gauge needle was inserted between the ligature site and the end of the cecum to extrude a small amount of cecal contents. The cecum was gently reduced and the laparotomy site was sutured. In the Sham group, animals underwent laparotomy and intestinal manipulation to expose the cecum without ligation or puncture. Monitoring was performed twice a day (once in the morning and once in the evening) and recorded every 4 h for 48 h after surgery. All mice received resuscitation through subcutaneous injection of normal saline at a 24 ml/kg body weight dose. Subsequently, all mice were anesthetized with pentobarbital sodium (40 mg/kg) and sacrificed for further analysis, including collection of serum as well as stomach and colon tissue. The animal experiments in the present study was conducted in the Liyang Hospital of Chinese Medicine and approved by the Experimental Animal Ethics Committee of Liyang Hospital of Chinese Medicine (Jiangsu, China; approval no. 2024LY-02-02-03).

The solvent DMSO was employed to dissolve atractylodin and prepare a storage solution. Prior to administration, the solution was diluted to the desired concentration using saline, ensuring that the final concentration of DMSO did not exceed 0.1%. The mice were randomly allocated into five groups: Sham operation group (Sham group), the cecal ligation perforation sepsis group (Model group), and the atractylodin treatment group (low, medium, and high doses), with 12 mice in each group.

One hour before surgery, the atractylodin treatment group received intraperitoneal injections of atractylodin at a concentration of 10 mg/kg/d for the low-dose group, a concentration of 20 mg/kg/d for medium-dose group, and a high-dose concentration of 40 mg/kg/d for high dose group. Post-surgery, mice received once-daily intraperitoneal injection doses of the designated atractylodin dese for 7 days. The dosage of the drug was obtained from the previous studies (18,19).

Clinical evaluation was performed with no pulsation on palpation of the carotid artery, detection of absence of corneal reflex, and secondary confirmation. In addition, the animals that died naturally were subjected to pathological analysis to exclude experimental interference factors. The present study strictly adhered to the following criteria to determine the timing of sacrifice through daily weight monitoring, a behavioral scoring system, and veterinary assessment, ensuring that animals did not suffer avoidable pain. Sacrifice were performed when one of the following situations occurred: i) Weight loss: A rapid weight loss of 15–20%; ii) weakness and loss of mobility: Unable to stand for more than 24 h, loss of appetite, dehydration; iii) infection and wound problems: A board-like abdomen upon abdominal palpation (indicating diffuse peritonitis); bloody discharge around the anus (indicating intestinal ischemic necrosis); iv) abnormal body temperature: A deviation of 4°C from the normal body temperature for more than 24 h; v) pain and behavioral abnormalities: Obvious signs of pain (such as aggression, aimless running), neurological symptoms (convulsions, paralysis).

A total of 90 C57BL/6 mice were used in the present study, of which six died naturally (autopsy showed sepsis), and the remaining mice were sacrificed at the end of the experiment. Isolfurane 4% (oxygen flow 2 l/min) was used for induction, and the oxygen flow rate was adjusted to 1.8% during the maintenance phase. The depth of anesthesia was verified by blood gas analysis after operation. Animals were sacrificed with an overdose of sodium pentobarbital (150 mg/kilogram) by intraperitoneal injection. Clinical evaluation was performed with no pulsation on palpation of the carotid artery, detection of absence of corneal reflex, and secondary confirmation. In addition, the animals that died naturally were subjected to pathological analysis to exclude experimental interference factors.

The stomach and colon tissue of the mice was fixed for 12–24 h at room temperature in 4% paraformaldehydeTissue embedded in paraffin was cut into 3 µm thick sections. The tissue sections were deparaffinized with xylene (twice, each for 5 min), followed by hydration with decreasing concentrations of ethanol. The sections were stained with hematoxylin for 15 min, then rinsed with tap water. The sections were then incubated in acid-alcohol for 30 sec, immersed in tap water for 15 min, and stained with eosin for 5 min. The sections were dehydrated with a gradient of ethanol (95, 95, 100, 100%, each for 2 sec) and cleared with xylene twice (each for 1 min). Finally, the sections were cleared with xylene, mounted with neutral resin, and observed under an inverted microscope.

After extracting total protein from the tissue using RIPA lysis buffer (Beyotime Institute of Biotechnology; cat. no. P0013), the BCA method is used to detect protein concentration. The protein extraction of stomach and colon tissues was boiled with gel-loading buffer for 10 min at 100°C, 50 µg protein was loaded per lane, and resolved by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and transferred to nitrocellulose membranes. The membrane was blocked with 5% BSA in 1X PBST buffer (0.01% Tween-20) for 90 min at room temperature and was probed with GAPDH) antibody (1:5,000; Proteintech Group, Inc., 10494-1-AP) or GPX4 antibody (1:1,000; Abcam, ab125066), Mitochondrial import receptor subunit TOM20 homolog (1:1,000; Abcam, ab186735), transferrin receptor protein 1 (TFR1) antibody (1:1,000; Abcam, ab214039), PRDX3 antibody (1:1,000; Abcam, ab73349), SIRT3 antibody (1:1,000; Proteintech Group, Inc., 10099-1-AP), SLC7A11 antibody (1:1,000; Proteintech Group, Inc., 32384-1-AP), occludin antibody (1:1,000; Abcam, ab216327), and zona occludens protein 1 (ZO-1) antibody (1:1,000; Proteintech Group, Inc.; cat. no. 21773-1-AP), for 2 h at room temperature. The membrane was washed three times (10 min each in 1X PBST) and incubated with secondary anti-rabbit IgG (H+L) antibody, (horseradish peroxidase conjugate) (1:5,000; Cell Signaling Technology, Inc., #7074), for 90 min at room temperature. After being washed three times (15 min each in 1X PBST), the membrane was incubated in a 2.0 mM DAB solution prepared in PBS for 5 min. The image of the immunoblot was digitalized with the GelDoc XR+ system and saved in TIF format. The protein levels were quantified by ImageJ software (National Institutes of Health, version 1.8.0) and normalized to GAPDH as an internal control.

For immunoprecipitation assays, Co-Immunoprecipitation kit (Beyotime Institute of Biotechnology; cat. no. P2175) was used following manufacturer instructions. Briefly, the tissue was collected and rinsed with PBS. Subsequently, tissue was lysed in the EBC buffer supplemented with protease inhibitors [50 mM tris (pH 7.5), 120 mM NaCl, and 0.5% NP-40]. Following ultrasonic lysis (power: 60%, ultrasonic intermittent time: 1 min, ultrasonic frequency: three times, ultrasonic exposure 10 s/time, 4°C), Protein A beads conjugated with anti-PRDX3 antibody (1:500, Abcam, cat. no. ab222807) were added to the lysate at a ratio of 20 µl bead suspension per 500 µl protein sample, followed by overnight immunoprecipitation at 4°C. The next day, the immunoprecipitate was pelleted by centrifugation at 500 × g for 3 min at 4°C. Then, the resulting product was washed in the lysis buffer. The immunoprecipitate was denatured at 100°C for 15 min with 50 µl 2X SDS protein loading buffer. The immunoprecipitate, input samples, and other lysates (10 µl) were separated by 10% SDS-PAGE and transferred to a PVDF) membrane for subsequent Western blotting. To detect acetylation of PRDX3, anti-Peroxiredoxin 3/PRDX3 antibody (Abcam; cat. no. ab222807) to precipitate PRDX3 from the samples. Subsequently, the Acetylated-lysine Antibody (Cell Signaling Technology, Inc.; cat. no. 9441) was used to assess the acetylated levels of PRDX3 through immunoimprinting.

DHE, a ROS-level indicative fluorescence probe (λex=535 nm, λem=610 nm), was used to detect intracellular superoxide anions. Fresh tissues are frozen (−20°C) and then cut into 10 µm thick sections using a DAKEWE 6250 cryostat microtome. The probe of DHE was incubated with tissue slices and the intensity of red fluorescence could reflect the level of ROS under a fluorescence microscope.

One Step TUNEL Apoptosis Assay Kit (Beyotime Institute of Biotechnology; cat. no. C1090) was used to detect apoptotic cells. After incubating with the TUNEL regent in the dark for 1 h at 37°C, cells were stained with DAPI (10 min at room temperature). Apoptotic cells showed red fluorescence. Cells were counted in three randomly selected fields of view using an inverted fluorescence microscope (20X magnification, Olympus IX83, Olympus Corporation).

H₂O₂ content was determined by H₂O₂ Content Detection Kit (Beijing Solarbio Science & Technology Co., Ltd.). The H₂O₂ detection reagent was melted in an ice bath. Next, the sample extraction solution or standard was added to the detection well, followed by the H₂O₂ detection reagent. This was gently mixed and allowed to rest at room temperature (15–30°C) for 5 min. A volume of 200 µl was aliquoted into a 96-well plate to measure the absorbance at 415 nm. Subsequently, the concentration of H₂O₂ in the sample was determined using a standard curve. Absorbance values were measured at 415 nm using a 96-well plate according to the manufacturer's instructions.

To clarify the effect of atractylodin on sepsis-induced mitochondrial oxidative damage, we detected the activity of catalase (CAT) in tissues using the Catalase (CAT) Activity Assay Kit (Solarbio, BC0205), determined the activity of superoxide dismutase 2 (SOD2) in tissues with the Superoxide Dismutase (SOD) Isoenzyme Activity Assay Kit (Solarbio, BC5255), and measured the content of glutathione/oxidized glutathione (GSH/GSSG) in tissues by means of the Total Glutathione (T-GSH)/Oxidized Glutathione (GSSG) Assay Kit (Nanjing Jiancheng Bioengineering Institute, A061-1-2), all in accordance with the instructions provided by the kit manufacturers. In addition, the contents of hydrogen peroxide (H₂O₂) and malondialdehyde (MDA) in tissues were detected using the Hydrogen Peroxide (H₂O₂) Content Assay Kit (Solarbio, BC3595) and the Malondialdehyde (MDA) Content Assay Kit (Solarbio, BC025), respectively.

The mitochondrial membrane potential was assessed utilizing an advanced mitochondrial membrane potential assay kit incorporating JC-1 (cat. no. C2003S; Beyotime Institute of Biotechnology) according to the manufacturer's instructions. In summary, mitochondria were isolated using a tissue mitochondria isolation kit in accordance with the manufacturer's protocol. The isolated mitochondria were subsequently incubated with 0.5 ml of JC-1 fluorescent dye for 30 min at 4°C in the absence of light, followed by analysis via flow cytometry (BECKMANCOULTER, CytoFLEX) using FlowJo software (v10.6.2, FlowJo, FlowJo Enterprise).

The activity of complex I–IV in mitochondrial was determined with the micro mitochondrial respiratory chain complex I, II, III or IV activity assay kit (Beijing Solarbio Science & Technology Co., Ltd.) according to the manufacturer's instructions. Briefly, stomach and colon tissues isolated from two respective mice in the same group were pooled and suspended in the mitochondrial complex extraction buffer, followed by gentle homogenization. The homogenate was centrifuged at 600 × g for 10 min at 4°C to remove cell debris and nuclei, and the supernatant was centrifuged again at 11,000 × g for 15 min at 4°C. The resultant pellet was resuspended in the extraction buffer and crushed by ultrasonication (power: 60%; intermittent time: 1 min; frequency: three times; 10 s/time, 4°C). The complex activity of mitochondrial homogenates in the respective reaction buffer was then measured spectrophotometrically at 340 nm (complex I), 605 nm (complex II), and 550 nm (complex III and IV), respectively.

Immunofluorescence staining was performed on paraffin-embedded tissue sections. Following dewaxing and hydration, antigen retrieval was performed using citrate buffer (pH 6.0), microwave treatment for 3 min. The sections are blocked with 1% goat serum (Beijing Solarbio Science & Technology Co., Ltd.; cat. no. SL038) in PBS at room temperature for 1 h and incubated with the primary antibody overnight at 4°C (Table SI. Subsequently, the sections are washed with PBST, incubated with goat anti-rabbit Alexa Fluor 647 secondary antibody (Thermo Fisher Scientific, Inc.) at a 1:400 dilution at room temperature for 1 h, followed by DAPI (10 µg/ml) staining for 10 min at room temperature, and mounting. Finally, the sections are observed under a fluorescence microscope (20X, Olympus, Tokyo, Japan) and quantified using. ImageJ software (National Institutes of Health, version 1.8.0).

TEM analyses were accomplished using HT7700 Hitachi Transmission Electron Microscope (Hitachi High-Technologies Corporation). TEM was used to observe the mitochondrial state of tissues. Freshly stomach and colon tissues were quickly cut into 1 mm cubes fixed overnight in 2.5% glutaraldehyde (4°C) and then post-fixed with 1% osmium tetroxide (4°C, 2 h), dehydrated through a graded ethanol series. Embedding was performed in a 1:1 resin (EMBed-812) and propylene oxide (Electron Microscopy Services) mix for 1 h at room temperature, followed by a 2:1 resin:propylene oxide mixture overnight at room temperature. Tissue were then placed in 100% resin for 3 h at room temperature. Ultrathin sections (70 nm) were collected and double-stained with uranyl acetate and lead citrate (room temperature for 15 min). Finally, the sections were visualized with a JEM1200ex Electron Microscope (JEOL).

AAV serotype 9 (AAV9)-short hairpin (sh)SIRT3 and AAV9-shNC were synthesized by Jikai Gene Chemical Technology Co., Ltd. Male WT C57Bl/6J mice (8-week-old) were injected in the tail vein with 2.5×10¹¹ viral genome particles of AAV 9. AAV9-mediated gene transfer and expression were allowed for 2 weeks before subsequent experiment.

Data are expressed as mean ± standard deviation. Comparisons among groups were tested with a one-way analysis of variance followed by the Tukey test. An unpaired t-test is used between the two groups (GraphPad Prism version 5; Dotmatics). P<0.05 was considered to indicate a statistically significant difference.

The present study initiated dose safety studies in mice to determine safety of atractylodin. C57BL/6 mice were treated with atractylodin (10, 20, 40 mg/kg/day) for 2 weeks to simulate long-term administration. Following common discovery-stage practices, multiple indices of liver function and renal function were examined. As shown in Fig. S1 atractylodin treatment did not markedly impair liver or renal function. The results showed that there was no significant difference in body weight, which proved the safety of the three groups of doses.

To evaluate the effects of atractylodin on acute GI, the present study created a model for sepsis-induced GI. The model of sepsis was induced by performing cecal ligation and perforation. Then, the corresponding treatments were performed in each group. The stomach and colon tissue morphology of each group were observed by H&E staining. The Sham group exhibited no evidence of intestinal mucosal injury in the pathological examination results. However, in the model group, there was an increase in epithelial space, a decrease in the number of villous epithelial cells and crypt cells, a disordered arrangement of villous epithelium, and some overflow of villous tips. Atractylodin treatment markedly attenuated sepsis-induced acute GI (Fig. 1A).

ZO-1 and Occludin are essential intestinal epithelial tight junction proteins. The present study investigated occludin and ZO-1 expression in the stomach and colon tissue with immunohistochemistry (Fig. 1B and E). The groups treated with atractylodin in low, medium and high doses appeared to relieve the pathological injuries to different degrees compared with the model group. The results demonstrated that the levels of ZO-1 and occludin in the model group were markedly lower than those in the Sham group, indicating impaired colon barrier function. Compared with the model group, the expression levels of ZO-1 and Occludin were markedly higher in all atractylodin treated groups (Fig. 1C, D, F and G).

The present study then assessed the role of atractylodin in modulating apoptosis of stomach and colon tissues in the mice using TUNEL staining. The number of TUNEL-positive cells was increased in mice's stomach and colon tissues in the model group, whereas atractylodin blocked this elevation (P<0.01; Fig. 1H-K). these results demonstrated that atractylodin represses the apoptosis of stomach and colon tissues in mice.

To further evaluate the effects of atractylodin on sepsis-induced GI mice, the level of pro-inflammatory cytokines were determined in the serum and tissues. The cytokines of IL-1β, IL-6, TNF-α, IL-4, and IL-10 were evaluated in the blood, stomach tissue, and colon tissue using ELISA (Fig. S2). Atractylodin effectively prevented the increase in pro-inflammatory cytokines, while increasing levels of anti-inflammatory cytokines IL-4, and IL-10.

Fig. 2A and B illustrates changes in antioxidant defense system indicators such as SOD2 and CAT levels in experimental mice. In the treatment group, there was a significant increase in CAT, SOD 2, and GSH/GSSG levels when compared with the model group in stomach and colon tissues. As shown in Fig. 2A-B, atractylodin treatment markedly decreased levels of H₂O₂ and MDA in a dose-dependent manner. In addition, mitochondrial morphology changes and function were evaluated by TEM. TEM revealed more disorganized, swollen, and damaged mitochondria in the model group. This mitochondrial damage was mitigated in the atractylodin-treated group in stomach and colon tissues (Fig. 2C). Semiquantitative western blotting analysis showed a trend to increase in TOM20/GAPDH ratio, which suggests an increase in mitochondrial number or mitochondrial dimensions in atractylodin-treated group compared with the model group (Fig. 2D-G).

The mitochondrial function was further examined. To assess the transmembrane potential, a flow cytometric-based assay was performed to measure transmembrane potential in each group in the stomach and colon tissue. Cells of the gastric and colon tissue of mice in each group were isolated and the mitochondrial membrane potential was detected using a mitochondrial membrane potential detection kit. We found that the potential for the dissipation of mitochondrial transmembrane could be attenuated by atractylodin in each group of stomach and colon tissue (Fig. 3A-D). To examine the ROS levels, stomach and colon tissue were stained with DHE, a superoxide indicator. The intensity of DHE fluorescence was markedly decreased in all atractylodin concentration groups compared with the model group (Fig. 3E-H). Mitochondrial complexes I, II, III, IV, and V activities were also markedly enhanced in each atractylodin concentration group compared with those in the model group (Fig. 3I and J).

To address the role of atractylodin in ferroptosis of sepsis-induced GI, the Fe²⁺ levels in stomach and colon tissues were determined. The results showed that the level of Fe²⁺ in stomach and colon tissues of mice was markedly decreased in the atractylodin-treated group (Fig. 4A and F). Based on the aforementioned results, the essential proteins related to ferroptosis were detected in each group. Western blotting showed that the levels of GPX4 and Solute carrier family 7 member 11 (SLC7A11) in the atractylodin-treated group were markedly higher compared with those of the model group. At the same time, TFR1 was decreased in both stomach tissues (Fig. 5B-E) and colon tissues (Fig. 5G-J). these results showed that atractylodin inhibited ferroptosis in mice with sepsis-induced GI.

It has been suggested that SIRT3-mediated deacetylation of PRDX3 could alleviate mitochondrial oxidative injury (11). The present study examined the expression of Ac-PRDX3 and SIRT3 in the stomach and colon tissues of mice. The expression of Ac-PRDX3/PRDX3 in the stomach tissues of mice in the atractylodin-treated group was lower than that in the model group. The high-dose atractylodin-treated group showed the lowest Ac-PRDX3/PRDX3 levels in stomach tissues (Fig. 5A and B). As shown in Fig. 5C and D, the SIRT3 expression markedly increased in the high-medium- and low-dose groups compared with the model group. The same trend was observed in the colon tissues (Fig. 5E-H). The expression of SIRT3 increased in the stomach and colon tissues of mice in the atractylodin-treated group compared with the model group (Fig. 5C, D, G and H). To explore how atractylodin regulates SIRT3 expression in cells, SIRT3 was colocalized with the nuclei marker DAPI using immunofluorescence in the stomach and colon tissues. As shown in Fig 5I and K, The increased SIRT3 expression with the atractylodin treatment groups was shown by immunofluorescence. The expression of SIRT3 increased with an increased dose of atractylodin in the stomach and colon tissues of mice in atractylodin treated group compared with the model group (Fig. 5J and L).

To assess whether atractylodin promotes the binding of SIRT3 to PRDX3, a co-immunoprecipitation assay was performed. The results showed that the binding of SIRT3 with PRDX3 was increased in all atractylodin treatment groups (Fig. 5M and N). The high doses of the atractylodin group showed increased SIRT3/PRDX3 and PRDX3/GAPDH in the stomach and colon tissues (Fig. 5N and O).

The downstream regulatory mechanism of the SIRT3 and PRDX3 was explored in sepsis-induced GI. To confirm the involvement of SIRT3 in the reparative effect on sepsis-induced GI, shSIRT3 was intravenously injected into the mice through the tail vein to knock down SIRT3. Fig. 6A and B showed that the protein levels of SIRT3 were markedly downregulated by 68% following injection with shSIRT3, as compared with the mice injected with shNC. Mice were divided into five groups: sham group, model group, model + atractylodin group, AAV-shSIRT3 group, and AAV-shSIRT3 + atractylodin group. The expression of GPX4 in the stomach and colon tissue was measured by eastern blotting. The results demonstrated that in the stomach and colon tissue, the GPX4 expression in the model group was markedly reduced compared with the sham group (Fig. 6C and E). However, it was obviously upregulated in the model mice treated with high-atractylodin. Compared with the model combined with atractylodin group, the GPX4 level of both the AAV-shSIRT3 and the AAV-shSIRT3 + atractylodin group was markedly decreased (Fig. 6D and F). These results suggested that SIRT3 plays a key role in ferroptosis-mediated GI. Western blotting also demonstrated that the protein expression of TOM20, ZO-1 and occludin was markedly decreased in the AAV-shSIRT3 + atractylodin treated group compared with the model combined with atractylodin group in the stomach (Fig. 6G-J) and colon tissues (Fig. 6K-N). These results showed that atractylodin inhibited mitochondrial oxidative stress by increasing SIRT3 expression. Moreover, mitochondrial complexes I, II/III, IV and V activities were also markedly attenuated in AAV-shSIRT3+ atractylodin-treated mice, as compared with model combined with atractylodin mice in the stomach (Fig. 6O-S) and colon tissues (Fig. 6T-X).