The extraction method of PAE using ethanol was referred from a previous study (Song et al., 2017). In this study, PAE was provided by the Department of Pharmaceutical Preparation, the First Affiliated Hospital of Guangzhou University of Chinese Medicine. PA (Yunfeng Periplaneta americana breeding base, Yunnan, China, Lot: 20210730) dried powder was extracted with 85% ethanol three times at 60°C. After standing overnight at 4°C, the upper layer of oil was removed. Petroleum ether was added to extract the active ingredient of PAE.

BALB/c male mice (16–20 g, 6–8 weeks old) were obtained from Guangdong Medical Laboratory Animal Center (License number: SCXK, Guangdong, 2022-0002). These animals were housed in the specific pathogen-free (SPF) animal laboratory of Guangzhou University of Chinese Medicine (License number: SYXK, Guangdong, 2018-0001). In this study, all protocols for animal experiments were approved by the ethics committee of Guangzhou University of Chinese Medicine (20220526013).

Mice were randomly assigned to six groups (Figure 1A): control, model, mesalazine, low-dose of PAE (PAE-L), high-dose of PAE (PAE-H), and DAPT (a Notch1 inhibitor) group. The UC model was established using 3% dextran sulfate sodium (DSS, MP Biomedicals, lot number: 160110) for 7 days. Mesalazine (520 mg/kg/day), PAE-L (100 mg/kg/day), PAE-H (200 mg/kg/day), and DAPT (10 mg/kg/day) were respectively administered via the intragastric route for 7 days.

In this experiment, the method of euthanasia of experimental animals is based on national standards of the Laboratory animal—Guidelines for euthanasia (GB/T 39760-2021). Mice were anesthetized by intraperitoneal injection of sodium pentobarbital once time (150 mg/kg). When the mouse’s heart stops beating, the colon and spleen tissue was collected for further study.

Disease activity index (DAI) = (weight loss percentage score + stool score + rectal bleed score)/3 (Guo et al., 2022). The occult blood test paper (Baso, Zhuhai, China, Lot: BA2020B) was assessed based on the timing of positive results using the Pyramidone Chemical Method (Supplementary Table S1).

The middle part of the colon (1 cm) was fixed in 4% paraformaldehyde solution for 24 h, embedded in paraffin, and cut into sections no larger than 0.3 mm thick. Hematoxylin and eosin (HE) staining was performed, and then pathological scores were determined by the severity of colonic mucosal injury on a scale of 0–4 according to Yan et al.’s report (Yan et al., 2022). The scoring table is shown in Supplementary Table S2.

Currently, due to the close relationship of basic principles, network pharmacology has been widely applied in deciphering the physiological activity mechanisms of TCM (Wang et al., 2021). We further predicted the targets and pathways for PAE intervention in UC using network pharmacological analysis. The chemical constituents of PA were searched by the Traditional Chinese Medicine Systems Pharmacology (TCMSP) database (http://tcmspw.com/tcmsp.php), Pubmed database (https://pubmed.ncbi.nlm.nih.gov/) and CNKI database (https://www.cnki.net/) using the keywords of “Periplaneta americana”, “Kangfuxinye” which is a patent medicine with PAE as its main component. Both the target of PAE chemical constituents and UC were obtained by the UniProt database (https://www.uniprot.org/) or CTD (Comparative Toxicogenomics Database, https://ctdbase.org/). The targets of PAE and UC were combined using the VENNY2.1 analysis tool (https://bioinfogp.cnb.csic.es/tools/venny/). The protein-interaction data of PAE-related targets and UC were obtained from the STRING database https://string-db.org/). In addition to the above, Cytoscape (https://cytoscape.org/) was used to build the Protein-Protein Interaction (PPI) and analyze the metabolic pathways of the related targets of PAE in the treatment of UC. Furthermore, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were used to analyze the metabolic pathways of the related targets of PAE in the treatment of UC by DAVID database (https://david.ncifcrf.gov/tools.jsp).

The spleen was prepared as single-cell suspensions, then stimulated and blocked with PMA (Lot: CS1001) and BFA (Lot: CS1002), which were purchased from MultiSciences (Lianke) Biotech, Hangzhou, China. The following antibodies were used for surface staining for 30 min at 4°C: FITC anti-CD4 (BioLegnd, San Diego, CA, United States, Lot: 100509), PE/Cyanine7 anti-CD25 (BioLegend, Lot: 102016). For intracellular staining with Alexa Fluor 647 anti-Foxp3 (BioLegend, Lot: 126408) and PE anti-IL-17A (BioLegend, Lot: 506904), cells were fixed and permeabilized with a True-Nuclear™ Transcription Factor Buffer Set (BioLegend, Lot: 424401). Samples were acquired and analyzed with a CytoFLEX LX flow cytometer (Beckman Coulter, California, United States) and the data was analyzed with CytExpert software. CD4⁺ CD25⁺ Foxp3⁺ cells as Treg cells were evaluated and CD4⁺ IL-17A⁺ cells were labeled as Th17 cells according to previous studies (Yu et al., 2015).

Tissues or cells were lysed with protein lysis buffer (KeyGEN BioTECH, KGP2100) containing proteinase and phosphatase inhibitors, and subsequently, the concentration of protein was quantified with BCA assay (KeyGEN BioTECH, KGP902). The equal amount of protein was separated with sodium dodecyl sulfate -polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membranes (Merck Millipore, LG18-1501M-1H). After blocking with 5% skim milk for 1.5 h (h) at room temperature, the membranes were incubated overnight at 4°C with different primary antibodies against Occludin (13409-1-AP, Proteintech), Claudin 1 (13050-1-A, Proteintech), ZO1 (TA5145F, Abmart; 21773-1-A, Proteintech), Foxp3 (Ab215206, Abcam), RORγt (13205-1-AP, Proteintech), Math1 (21215-1-A, Proteintech), Notch1 (T55244S, Abmart; 20687-1-AP, Proteintech), Hes1(11988S, CST) and β-actin (ab179467; Abcam). Followed by secondary antibodies for 1 h. The PVDF membranes were detected using a chemiluminescence imager (Bio-Rad) and quantified using the Image Lab software (Version 4.1, Bio-Rad).

As described in immunohistochemistry (IHC) immunofluorescence (IF) by Xiao et al. (2024), the blank sections were dewaxed, antigen retrieval in turn. After blocking with 5% bovine serum albumin (BSA), the primary antibody was incubated at 4°C overnight, and the secondary antibody was incubated for 30 min at 37°C. Finally, diaminobenzidine (DAB) was used to stained IHC samples, or IF with DAPI.

Jurkat cells (Pricella, CL-0129) were cultured in RPMI-1640 medium with 10% fetal bovine serum (FBS) at 5% CO₂, 37°C (Yang et al., 2022). The cells were treated with the Notch1 agonist valproic acid (VPA. Selleck, S1168) in different concentrations (0.4, 0.8, 1.6, 3.2, 6.4, 12.8 mM) for 24, 48 h in 96 wells-plates. Next, CCK-8 reagent (GLPBIO, GK10001) was added into each well for 4 h. The optical density values were measured at 450 nm using a microplate reader. For more details about PAE (0.01, 0.1, 1, 10 mg/mL) processing, refer above.

Jurkat T cells were treated with PBS, VPA (3.2 mM VPA), and PAE (3.2 mM VPA+1 mg/mL PAE) for 24 h. Then notch1-activated Jurkat T cells were co-cultured with HCoEpic cells for 10 h.

The results are the mean ± standard deviation. Data were analyzed by SPSS 19.0 (IBM, Armonk, NY, United States). One-way analysis of variance was used for the comparison of data between groups. P < 0.05 was considered statistically significant.

Compared with the control group, the DAI score of the model group was significantly higher on the second day with 3% DSS treatment. Compared with the model group, lower DAI scores were observed in the mesalazine, PAE-L, PAE-H, and DAPT groups (Figure 1B). As illustrated in Figures 1C, D, the colon of the model group was obviously wrinkled, and the unit weight of colon in the model group (colon weight/length) was increased compared with the control group (P < 0.01). Treatment with mesalazine, PAE-L, PAE-H and DAPT significantly increased the colon length in UC mice (P < 0.05, P < 0.01). In addition, there was a significant decrease in the unit weight of the DAPT group compared to the model group (P < 0.05), while there was no statistical significance in the mesalazine and PAE groups (P > 0.05).

HE Staining showed intestinal barrier damage in the DSS-included UC mice, including disordered mucosal gland structure and diminished glands and goblet cells. In addition, a large number of inflammatory cells infiltrated in the mucosal and submucosal layers, individually involving the muscular layer. Damage to the colon intestinal mucosa was improved with the treatment of PAE-L, PAE-H, and DAPT (Figures 1E, F).

Tight junctions (TJs) are essential for the formation and maintenance of the intestinal barrier by enhancing epithelial cell-cell adhesion (Balda and Matter, 2023). The levels of tight junction proteins, including occludin, claudin1, and ZO1 were further measured to evaluate the injury of colonic mucosa in UC mice. It was found different degrees of upregulation of occludin, claudin1, and ZO1 levels in UC mice with PAE, mesalazine, and DAPT treatment (P < 0.05, P < 0.01), suggesting that PAE and DAPT could protect the colonic mucosal barrier by increasing the expression of occludin, claudin1, and ZO1 (Figures 1G–J).

A total 4 components were identified in PAE by high-performance liquid chromatography (HPLC), including uracil (C₄H₄N₂O₂), hypoxanthine (C₅H₄N₄O), xanthine (C₅H₄N₄O₂) and inosine (C₁₀H₁₂N₄O₅) (Supplementary Figure S1; Supplementary Table S3). In addition, cytosine (C₄H₅N₃O) and cytidine (C₉H₁₃N₃O₅) were supplemented through PubMed and CNKI databases. A total of 31 targets of PA were identified from the TCMSP database, and 29, 051 disease targets were yielded in the CTD database. After overlapping of common targets with Venn diagrams, 24 common targets of the active ingredient of PA targeting UC were obtained (Figure 2A). Furthermore, the PPI network showed the top ten targets including IL-6, IL-10, IL-1B, IL-17A, STAT3, IL-4, IFNG, FOXP3, MMP9, and TGFβ1 (Figure 2B).

GO enrichment identified that UC intervention by PA may be related to the positive regulation of IL-6 production, negative regulation of cell proliferation, and factor activity (Figure 2C). KEGG enrichment analysis showed that the targets of UC intervention by PA may be closely related to Th17 cell differentiation, the IL-17 signaling pathway, and cytokine-cytokine receptor interaction (Figure 2D).

The percentages of Th17 and Treg cells were detected by flow cytometry assay. The results showed that the percentage of Th17 cells was increased, while the percentage of Treg cells was remarkably decreased in the model group compared with the control group. After treatment with PAE, or DAPT, the percentage of Th17 cells was obviously reduced, and the percentage of Treg was increased compared with the model groups (P < 0.01) (Figure 3A).

RORγt and Foxp3 are the critical transcription factors for the development and function of Th17 and Treg cells, respectively. Furthermore, IL-17A and TGF-β are crucial Th17 and Treg cell effectors. As shown in Figures 3B–F, RORγt and IL-17A were significantly increased in the colon of the model group compared to the control (P < 0.05), while mesalazine, PAE-L, PAE-H and DAPT treatment dramatically decreased the expression of RORγt and IL-17A (P < 0.05, P < 0.01). TGF-β downregulated the expressions in the model group compared to the control (P < 0.05), while PAE and DAPT intervention could upregulated the expressions (P < 0.01). In addition, there was no statistical difference in Foxp3 expression between the groups (P > 0.05).

As seen in Figure 4, the protein expression of Notch1 was increased, and the Math1 was reduced in the colon of the model group compared with the control group. Conversely, Notch1 expression was downregulated and Math1 expression was upregulated in the PAE-H and DAPT groups compared with the model group (P < 0.05). It suggested that PAE could regulate DSS-induced Notch signaling activation in UC mice.

To identify a possible effect of VPA (a Notch1 agonist) and PAE in vitro, the proliferation of Jurkat T cells was analyzed with CCK-8 assay. It indicated that cell viability was greater than 80% at 0.4–1.6 mM VPA and highest at 0.8 mM for 24 h (Figure 5A). PAE (0.1 mg/mL and 1 mg/mL) administration significantly promoted cell proliferation for 24 h. The overall cell viability of different concentrations of PAE at 24 h was higher than that at 48 and 72 h (Figure 5B). To ensure cell viability as much as possible, we built the model using VPA (0.8 mM) and PAE (1 mg/mL) for 24 h.

VPA obviously upregulated the expressions of Notch1 and downregulated Math1 (P < 0.05). Meanwhile, the ratio of RORγt/Foxp3 was upregulated in the VPA group (P < 0.05). PAE intervention could alleviate this situation compared with the VPA group (P < 0.05) (Figures 5C–G).

To further determine the relationship between Notch1 and Th17/Treg in the colonic mucosal damage, Notch1-activated Jurkat T cells were co-cultured with HCoEpic cells. The results showed that the expressions of Occludin, ZO1 was higher in the PAE group compared with VPAs (P < 0.05) (Figures 5H–K).