Human colon cell lines (HCT116 and SW480) were obtained from the American Type Culture Collection (ATCC, Rockville, MD). All cell lines were incubated in RPMI-1640 (Invitrogen) and DMEM (Invitrogen), supplemented with 10% (v/v) fetal bovine serum (Invitrogen) and 1% (v/v) penicillin–streptomycin (Invitrogen) at 37°C in a humidified atmosphere of 5% CO₂. C57BL/6J male mice (5–6 weeks old, 18–22 g) were purchased from the the Animal Supplier Center of Binzhou Medical University. All procedures involving laboratory animals were in accordance with the guidelines of the Instituted Animal Care and Use Committee of Binzhou Medical University. All protocols were submitted and validated by the Animal Care Ethics Committee of Binzhou Medical University. Bone marrow–derived macrophage (BMDM) cells were isolated according to the following procedures. Bone marrow cells were isolated from C57/BL6 mice and cultured with DMEM supplemented with 10% fetal bovine serum and 20 ng/ml GM-CSF (Peprotech, Rock Hill, NJ). The culture fluid was exchanged to a fresh culture medium every 3 days. Under these conditions, adherent macrophages were obtained within 7–8 days. Cells were harvested and seeded on 24-well plates. After culture for 6 h without GM-CSF, the cells were used for the experiments as bone marrow–derived macrophages.

Atractylenolide I was obtained from Chengdu Herbpurify CO. Ltd. (Chengdu, China), dissolved in dimethylsulfoxide (DMSO; Sigma-Aldrich, St. Louis, MO, United States), at 100 mM stock and diluted immediately before each experiment.

Cell viability assays were performed using the Cell Counting Kit-8 (CCK-8; Dojindo Molecular Technologies, Japan). The cells were inoculated in a 96-well plate with culture medium and incubated at 37°C for 24 h. After acclimatization, the cells were treated with atractylenolide I for 24 and 48 h. Then the medium was replaced with fresh medium containing 10 ml of the CCK-8 solution. After incubating for 2 h at 37°C, the optical density (OD) at 490 nm was measured.

Cells were treated with atractylenolide I for 24 h, and then the cells were processed by ROS analysis. In brief, 5 × 10⁴ cells were suspended in 10 μmol/L of DCFH-DA solution and incubated for 30 min in the dark according to the manufacturer’s instructions. The ROS level was analyzed by FACS flow cytometry.

Cells were stained with a Annexin V-FITC Apoptosis Detection Kit (BD Biosciences, San Jose, CA, United States). According to the manufacturer’s instructions, the cells were incubated with 5 ml of Annexin V and 5 ml of propidium iodide (PI) for 15 min at room temperature, and then the stained cells were analyzed on a FACS flow cytometer.

Cells were treated with atractylenolide I for 24 h, and then the cells were performed by a mitochondrial membrane potential (MMP) assay. In brief, cells were suspended in 2 μmol/L of JC-1 solution and incubated for 30 min in the dark according to the manufacturer’s instructions. The MMP level was analyzed by FACS flow cytometry.

Protein expression was analyzed by Western blot. The total protein (50 μg) was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. After the protein is transferred to the polyvinylidene fluoride microporous membrane (Bio-Rad), the membrane is blocked with 5% skimmed milk powder and the primary antibody [anti-Bax, anti-PCNA, anti–β-catenin, anti-Drp1 (1:1,000 dilution, American Abcam), anti-NLRP3 (1:500 dilution, American Abcam), anti-ASC, or anti–caspase-1]; then anti-mouse (anti-rabbit (IgG)) is linked with fluorescein (1: 1,000), and then incubated with an antibody conjugated to anti-fluorescein alkaline phosphatase (1:5,000). The immune complexes were detected with enhanced chemiluminescence reagents. For quantification, the signal optical density was normalized to β-actin by Quantity One image analysis software.

The animals were kept under controlled conditions in a regular animal community (22°C, 12 h/12 h dark/light cycle). At the beginning of the experiment, the mice were between 5 and 8 weeks old and weighed 15–22 g. 46 mice were used in the study, 10 mice were in the normal control group (control group), 20 mice were in the AOM/DSS (model) group, and 16 mice were in the AOM/DSS and atractylenolide I group. At the end of the study, there were 10 surviving mice in the control group, 12 surviving mice in the model group, and 12 surviving mice in the atractylenolide I group. DSS is shown in Figure 3A. On day 1, C57BL/6J mice were injected with AOM (10 mg·kg-1, i.p.). After 5 days, 4% DSS (International Laboratory, Chicago, Illinois, United States) was added to drinking water for 7 days, and then tap water was added for another 14 days for recovery. This cycle was repeated twice. Atractylenolide I (25 mg/kg, 50 mg/kg) or the equivalent concentration of normal saline was added twice a day with a gavage tube. At the 10th week, the mice were killed by cervical dislocation, and the colon was removed (from the ileocecal junction to the edge of the anus). Record the number, size, and location of preneoplastic and neoplastic lesions (dysplasia and cancer) in the colon according to the gross examination. Measure the size of dysplasia by a visual micrometer. After a rough inspection, the colon was cut into small pieces at approximately 1 cm intervals and stored at −80°C for further immunohistochemistry and protein analysis.

For histopathological examination, the formalin-fixed and paraffin-embedded colon tissue was cut into serial sections (5 μm) and stained with hematoxylin and eosin (H& E). Histological changes were confirmed, such as dysplasia and cancer. The assessment of dysplasia and adenocarcinoma was based on the criteria proposed in the mice Models of Intestinal Cancers consensus report. The height of the intestinal villi is the distance from the base of the villi (the junction between the intestinal glands and the villi) to the top of the villi. Intestinal villi were measured using the Image-pro plus 6.0 image analysis software.

The paraffin-embedded colon sections were deparaffinized, rehydrated, and pretreated with hydrogen peroxide in a PBS buffer. Heat-induced antigen retrieval was performed. The sections were incubated with anti-Drp1 (1:1,000 dilution, Abcam, United States) and anti-NLRP3 (1:100 dilution, Abcam, United States). The sections were incubated with HRP-conjugated secondary antibody, tyramide amplification was performed first, streptavidin-HRP was added, and then the DAB kit was used to show a positive signal. The sections were inspected at 400× magnification and analyzed using NIS-Elements. The following formula was used to calculate the positive content: positive content (PC) = average optical density x positive area.

The mice in the control group and the treatment group were sacrificed, and the mice serum was collected. The concentration of IL-1β in serum was determined by ELISA (Ebiosciences-Easy-Set-Go) according to the manufacturer’s protocol.

The pLVX-IRES-ZsGreen1 lentivirus vector was purchased from Genomeditech (Shanghai, China). To clone human Drp1 cDNA into the pLVX-IRES-ZsGreen1 vector, XhoI and NotI restriction sites were inserted into the PCR products. After confirming the Drp1 sequences by sequencing, the plasmid was cotransfected into 293T cells with the lentivirus packaging plasmids pMD2.G and psPAX2 to produce lentivirus particles. The Drp1 lentivirus particles were then used to infect mice BMDMs, and the control and AOM/DSS mice models. Four weeks after the mice were modeled, Drp1-overexpressed lentivirus was injected into the tail vein. The lentivirus titer was 1 × 10⁸, and the injection dose was 100 μl, it was injected only once. The mice model and BMDMs stably overexpressing Drp1 or empty vector were established by first transducing the cells with Drp1 lentivirus particles or empty vector lentivirus particles.

To determine the production of mROS, cells were stained with MitoSOX (Invitrogen) according to the manufacturer’s protocol. Then the fluorescence of the cells was monitored by a flow cytometer (FACSVerse, BD).

Each experiment was repeated at least three times. Data were expressed as mean ± SD. All data were analyzed using SPSS statistical software package (version 23.0, SPSS Inc., Chicago, Illinois, United States). The data were compared between the two groups with two independent sample tests. The average of data from more than three groups was compared with one-way analysis of variance (ANOVA), and multiple comparisons were made. A value of p < 0.05 was considered statistically significant.

To address the role of atractylenolide I on the cell viability, HCT116 and SW480 (CRC cell lines) were treated with different concentrations of atractylenolide I (0–200 μg/ml). As shown in Figure 1, atractylenolide I significantly reduced the viability of HCT116 cells in a dose- and time-dependent manner with an IC50 value of 630 (atractylenolide I) μg/mLfor 24 h (Figure 1A). The results also indicated that 200 μg/ml of atractylenolide I was weakly cytotoxic (10–15% inhibition rate) for SW480 cell (Figure 1B). Thus, HCT116 cells were used in subsequent experiments to study the mechanism of atractylenolide I. Annexin V/propidium iodide (PI) double staining (Figure 1C) showed that most of the cell deaths in HCT116 induced by atractylenolide I could be attributed to apoptosis. But in terms of inhibiting cell proliferation, the ability of atractylenolide I was a relatively low degree.

We detected the changes in ROS levels in HCT-116 cells treated with atractylenolide I (25, 50, and 100 μg/ml) by loading DCFH-DA fluorescent probes. The analysis of the results of the fluorescence intensity of DCF detected by flow cytometry showed that compared with the control group, the intracellular ROS content in the atractylenolide I groups was significantly increased (Figure 1D).

We used the JC-1 method to detect the integrity of the mitochondrial membrane, where red and green fluorescence represent the normal mitochondria and the damaged mitochondria, respectively. Figures 2A,B showed a dose-dependent decrease in the ratio of red/green fluorescence in the cells treated with atractylenolide I, indicating that atractylenolide I could induce mitochondrial dysfunction and apoptosis. Atractylenolide I increased the expression of Bax in HCT116 cells. Reduction of Drp1 was observed by atractylenolide I in HCT116 cells (Figures 2C,D). Combined with the previous research in CRC cells, the dissipation of ΔΨm and the activation of Bax and downregulation of Drp1 indicated that the cell death induced by atractylenolide I could be regarded as intrinsic apoptosis, and atractylenolide I might inhibit the Drp1-mediated mitochondrial fission.

AOM/DSS was used to construct the mice model of caCRC, and the effect of atractylenolide I on caCRC was observed (Figure 3A). The survival rate was observed in mice exposed to AOM/DSS (Figure 3B). Atractylenolide I significantly delayed mass death in the AOM/DSS mice model (p-value 0.0338). A dose gradient of atractylenolide I (25 mg/kg, 50 mg/kg) was prescreened. The body weight fluctuation was substantially alleviated in mice receiving atractylenolide I treatment (25 mg/kg/d: 23.76 ± 0.11, p-value 0.0028; 50 mg/kg/d: 24.29 ± 0.25, p-value 0.0000) (Figure 3C). As shown by the survival rate and body weight of the mice, atractylenolide I was well tolerated in mice, and no obvious systemic toxicity was observed during the entire course of drug treatment. Most neoplasms were disseminated in the middle and distal colon, and were confirmed histologically (Figure 3D). Atractylenolide I treatment significantly reduced the number of large tumors (>2 mm). The number of neoplasms larger than 2 mm in atractylenolide I treatment group was 30.77% less than that in the untreated mice (Figure 3E).

H&E staining showed congestion of intestinal mucosa, massive loss of surface epithelium, and unclear colonic villi in the AOM/DSS mice model group. Colonic villi were abundant in the atractylenolide I group, and the height was significantly increased (Figure 4A). The NLRP3 inflammasomes have been shown to play a key role in AOM/DSS-induced caCRC model (Dai et al., 2020). To determine whether atractylenolide I inhibits NLRP3 inflammasomes in vivo, we evaluated the expression of NLRP3 in the AOM/DSS mice model by immunohistochemistry (IHC) and Western blotting. Atractylenolide I similarly reduced the increased secretion of IL-1β in serum (Figure 4B). In addition, atractylenolide I significantly reduced the accumulation of NLRP3 in the intestine (Figure 4C). Atractylenolide I also significantly reduced the expression of intestinal NLRP3, caspase-1, and ASC induced by AOM/DSS (Figure 4D). Overall, the results showed that atractylenolide I inhibited the activated NLRP3 inflammasomes in the AOM/DSS mice model.

The mitochondrial dysfunction was further confirmed by the decreased mitochondrial membrane potential and the expression level of Drp1 proteins of atractylenolide I–treated HCT116 cells. Atractylenolide I significantly reduced the expression of mitochondrial fission marker, Drp1, in the AOM/DSS mice model by Western blotting (Figure 4D) and immunohistochemistry (Figure 5A). The immunoblotting assessment (Figure 5B) showed that the cleavage of poly (ADP) ribose polymerase (PARP), caspase-3, and Bax was also increased in the AOM/DSS mice model, which were treated with atractylenolide I. In addition, the morphology of the mitochondria detected by TEM showed a weak mitochondrial fission ability in a atractylenolide I-treated AOM/DSS mice model (Figure 5C), suggesting that atractylenolide I inhibited mitochondrial fission to induce apoptosis in the AOM/DSS mice model.

In order to determine whether Drp1 was involved in the activation of the NLRP3 inflammasomes, the overexpression of Drp1 in control and AOM/DSS mice model transduced with the Drp1 lentivirus vector were established (Figure 6A). We found that the body weight fluctuation was not substantially alleviated in mice receiving atractylenolide I treatment, which were transduced with the Drp1 lentivirus vector (ATR I + LV-EV: 24.32 ± 0.18 vs ATR I + LV-Drp1: 23.04 ± 0.47, p-value 0.0000) (Figure 6B). In addition, overexpressing the Drp1 expression counteracted the restraint of atractylenolide I on the release of IL-1β (Figure 6C). Furthermore, transducing with the Drp1 lentivirus vector significantly upregulated β-catenin and PCNA expression in the atractylenolide I–treated AOM/DSS mice model (Figures 6D,E). Upregulation of Drp1 reversed the inhibitory effect of atractylenolide I on the activation of NLRP3 inflammasomes (Figures 6D,E). Collectively, these results indicated that Drp1 was related to the regulation of atractylenolide I on the activation of NLRP3 inflammasomes in the AOM/DSS mice model.

In order to gain insights into the mechanism of the anti-inflammatory effect of atractylenolide I, we studied the effect of atractylenolide I on the activation of NLRP3 inflammasomes in vitro. As shown in Figure 7A, atractylenolide I showed a significant dose-dependent inhibitory effect on the activation of IL-1β in BMDMs stimulated by LPS/DSS. LPS/DSS-stimulated secretion of IL-1β was inhibited by Drp1 inhibitor Mdivi-1, as well as MCC950, a potent and selective small molecule inhibitor, and can normally and atypically inhibit NLRP3 activation. Also, atractylenolide I significantly reduced the expression of Drp1 in BMDMs stimulated by LPS/DSS (Figure 7B). Mdivi-1 treatment reduced the expression of NLRP3, ASC, and pro–caspase-1 in BMDMs stimulated by LPS/DSS (Figure 7C). To generate macrophages with the overexpression of Drp1, the Drp1 lentivirus (LV-Drp1) expression vector transduction and empty vector (EV) transduction were used to infect BMDMs. At the protein immunoprecipitated with anti-ASC, atractylenolide I greatly inhibited the expression of NLRP3 and caspase-1, but it had no effect on the expression of NLRP3 and caspase-1 in atractylenolide I–treated Drp1 overexpression BMDMs (Figure 7D). Overexpressing Drp1 expression counteracted the restraint of atractylenolide I on the release of IL-1β of LPS/DSS-stimulated BMDMs (Figure 7E). Taken together, these data indicated that atractylenolide I could inhibit the NLRP3 expression and subsequent IL-1β secretion induced by NLRP3 inflammasome activation in a Drp1-dependent manner.

We next detected the effect of atractylenolide I and MCC950 on the mitochondrial ROS level. Atractylenolide I and MCC950 decreased the level of mtROS induced by the addition of LPS/DSS in vitro. Of more interest, after LPS/DSS stimulation activated NLRP3, the level of mtROS in Drp1 overexpression BMDM was significantly higher than that in control BMDM (Figure 7F). These findings increased the possibility that Drp1-mediated mitochondrial fission might play an effective role in NLRP3 inflammasome activation.