Data collection and preprocessing

The data used for our analysis were obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). We selected gene expression profiles from GEO datasets (GSE134025, GSE42911, GSE87473) that composed of 124 samples in total. The expression data were normalized and integrated by the “limma” and “sva” package in R. Principal Component Analysis (PCA) was utilized to enhance the accuracy of different sample data.

Differential expression analysis

“limma”, “dplyr”, “pheatmap”, “ggplot2” packages in R were utilized for differential expression analysis of gene expression data. The filtering criteria for differentially expressed genes was set as |log2FC|> 1 and p-value < 0.05. The genes related to cellular senescence were obtained from the database of cell senescence genes (http://genomics.senescence.info/cells/). “VennDiagram” was applied to draw Venn diagrams displayed the intersecting genes between differentially expressed genes in UC and cellular senescence genes.

Functional enrichment analysis

To elaborate understand the roles of the intersecting genes in different biological processes, we performed Gene Ontology (GO) enrichment analysis using the “clusterProfiler”, “org.Hs.eg.db”, “enrichplot”, “ggplot2”, “GOplot”, “circlize”, “RColorBrewer”, “ComplexHeatmap” R packages. “clusterProfiler”, “org.Hs.eg.db”, “enrichplot”, “ggplot2” R packages was used for KEGG enrichment analysis to determine which pathways these genes are significantly enriched in.

Protein–protein interaction

We conducted protein–protein interaction (PPI) analysis on the intersecting genes using the STRING website (http://cn.string-db.org/), excluding unbound proteins from downstream analysis. Cytoscape software was employed for visualizing the PPI networks, with upregulated and downregulated genes distinguished by different colors. The MCODE plugin in Cytoscape was used to generate subnetworks, while the CytoHubba plugin and the UpSetR package were utilized to identify and visualize hub genes within these networks. Additionally, we constructed networks of hub genes using the GeneMANIA database (http://genemania.org/). A Perl script was employed to predict miRNAs that bind to the hub genes. The spongSPAN tool and Perl scripts were used to identify IncRNAs that interact with these miRNAs. The ceRNA regulatory network of the hub genes was constructed using Cytoscape software. The TRRUST database (http://www.grnpedia.org/trust/) and a Perl script were utilized for the enrichment analysis of transcription factors associated with the hub genes and for constructing the transcription factor regulatory network. Finally, Cytoscape software was used to visualize the transcription factor regulatory network of the hub genes.

Drug enrichment analysis

DSigDB database (http://dsigdb.tanlab.org/) was utilized to obtain the relationships between drugs and genes. “clusterProfiler”, “org.Hs.eg.db”, “enrichplot”, “ggplot2” packages in R were applied to drug enrichment analysis. To conduct molecular docking, we downloaded the 3D conformers of the drugs from the PubChem database (http://pubchem.ncbi.nih.gov/) and protein structure from RCSB database (http://www.rcsb.org/). CB-DOCK2 website (http://cadd.labshare.cn/cb-dock2/index.php) was employed to generate protein–ligand docking diagram.

Cell culture and treatment

The human colon adenocarcinoma cell line Caco-2 was obtained from the ATCC, the experimental cells were first maintained in DMEM/F12 medium supplemented with 10% FBS at 37℃ and 5% CO2. For the LPS (MedChemExpress) group, Caco-2 cells were stimulated with LPS (1 μg/ml) for 24 h.

Quantitative real-time PCR (qRT-PCR) analysis

Total RNA from tissue cultured cells was extracted using TRIzol reagent (Thermo Fisher). Quantitative real-time PCR (qRT-PCR) was conducted using SYBR qPCR Master Mix (Vazyme), with GAPDH as the internal control, primer oligonucleotide sequences are as follows. To delineate the mechanistic pathway, P21 and P16 expression levels were assessed after LPS stimulation in the presence or absence of either the LCN2 inhibitor ZINC00640089 (MedChemExpress) or 4-HPR (MedChemExpress). qRT-PCR was also employed to verify the overexpression of LCN2 in CD4+ cells.

Western blot

To detect the expression level of LCN2 in UC, the tissue samples from four UC patients were collected to detect the expression of LCN2. The intestinal tissues of mice in the in vivo experiments were used to detect the expression of P16, P21, and LCN2. The tissues were lysed with RIPA buffer containing protease inhibitor and phosphatase inhibitor. After BCA quantification, the samples were loaded on SDS-PAGE gels, electrophoresed, transferred onto PVDF membranes, blocked with 5% BSA and incubated with the primary antibodies overnight at 4℃. The next day, the membranes were washed with TBST and incubated with a secondary antibody (Cell Signaling Technology) one hour at the room temperature.

In vivo studies

Animal experiments were performed in the Laboratory Animal Center of Peking University First Hospital. We performed a dose-ranging study of 4-HPR in a dextran sulfate sodium (DSS, MP Biochemica)-induced acute colitis mouse model. Healthy male C57BL/6 J mice aged 8 weeks (n = 5) were used as experimental mice in vivo studies. We randomly assigned mice into five groups, which respectively received the following treatments for 7 consecutive days: normal drinking water (Control), 3% DSS drinking water (DSS), 3% DSS drinking water + 4-HPR (25 mg/kg), 3% DSS drinking water + 4-HPR (50 mg/kg), and 3% DSS drinking water + 4-HPR (100 mg/kg). 4-HPR was intraperitoneally injected every day in treated group. During the 7-day experimental period, we continuously monitored body weight and inflammatory scores in mice, with colon length measurements obtained following sacrifice at study termination. Following dose optimization, 4-HPR was re-administered in the murine acute colitis model, with consistent evaluation criteria. All mice were divided into three groups: Control group, DSS group, DSS with 4-HPR (50 mg/kg) treated group. During the whole experimental period, disease progression was systematically evaluated through daily assessment of colitis severity markers: Body weight variation (expressed as percentage change from baseline); Stool characteristics (scored 0–3: 0 = formed pellets, 3 = liquid diarrhea); Fecal occult blood (graded 0–3: 0 = negative, 3 = gross bleeding). Peripheral blood was collected from the mice prior to euthanasia, and biochemical indicators (AST, ALT, BUN, CREA) were measured for safety assessment.

Assessment of Intestinal Permeability

Intestinal permeability was evaluated using fluorescein isothiocyanate-dextran 4 (FD4, Sigma-Aldrich). Briefly, mice were fasted for 12 h before oral gavage of FD4 (500 mg/kg), diluted in PBS. After 4 h, Blood from the eyeballs was collected in the anticoagulant tube. FD4 levels in serum were quantified using a fluorescence microplate reader at excitation/emission wavelengths of 492/520 nm. A standard curve was generated using serially diluted FD4 to calculate sample concentrations. In addition, LPS levels in the serum were determined by mouse lipopolysaccharides (LPS) ELISA kit (CUSABIO).

Histopathology analysis

The colon tissue was fixed in 4% paraformaldehyde for 48 h, followed by paraffin embedding and sectioning at 5 μm thickness. The sections were deparaffinized and subsequently stained with hematoxylin and eosin (H&E) for histological observation under a microscope, with representative fields of view selected for analysis. We evaluated histopathological changes using the scoring criteria [10]. The scoring system evaluated from three pathological parameters: crypt architecture (0: normal structure; 1: slight loss of crypt fossae; 2: majority of crypt fossae missing; 3: severe crypt fossae loss), inflammatory cell infiltration (0: normal tissue architecture; 1: mild infiltration; 2: moderate infiltration; 3: severe/dense infiltration), and muscle layer thickening (0: normal muscularis propria; 1: mild hypertrophy; 2: moderate hypertrophy; 3: marked hypertrophy). Each parameter is scored independently on a 4-point scale (0–3). The total histopathological score was calculated by summing the individual scores of three parameters.

Immunofluorescence

To explore the Intestinal epithelial barrier function, the intestinal tissues were embedded and sectioned. These slides were deparaffinized and rehydrated, antigens were retrieved using pH 9.0 Tris–EDTA buffer. Slides were blocked and incubated with rabbit anti-mouse ZO-1 and Occludin antibody (Thermo Fisher) overnight at 4℃. The slides were washed and incubated with secondary antibodies Goat anti-rabbit Alex Fluor 488 and 647 (Thermo Fisher) and the nuclear counter stain was DAPI (Thermo Fisher).

SA-β-Gal staining

Fresh colon tissues were processed into cryosections at 8–10 μm thicknesses. Tissues were fixed with β-galactosidase staining fixative at room temperature for 20 min, followed by three 5-min PBS washes. Sections were incubated with sufficient β-galactosidase solution at 37 °C in the dark overnight. Then sections were washed twice with PBS and twice with distilled water. Nuclear counterstaining was performed using Nuclear Fast Red (G1035, Servicebio) for 3 min, followed by three water rinses. Dehydration was achieved via two 5-min ethanol (100%) incubations, cleared in xylene (5 min), and mounted with neutral resin. Senescent cells were identified by cytoplasmic blue puncta.

Flow cytometry

The spleens of the mice were ground and then red blood cell lysis buffers were added and centrifuged at 300 × g for 10 min. The isolated cells were incubated with anti-CD3-APC (BioLegend), anti-CD4-PE (BioLegend), anti-CD45-FITC (BioLegend), anti-FOXP3-BV421 (BioLegend), anti-RORγt-BV650 (BD Biosciences) and 7AAD. The distribution of Th17 and Treg cells in CD4+ T cells was calculated.

Enzyme-linked immunosorbent assay (ELISA)

IL-1β, IL-6, TNF-a, IL-10 and IL-17A were measured using cytokine-specific ELISA kits (Elabscience). The assay was performed according to the manufacturer’s protocol with minor modifications. Briefly, standards, blanks, and samples were added to a pre-coated microplate and incubated for 90 min at 37 °C. Then 100 µL of biotinylated detection antibody working solution was added, followed by a 1-h incubation at 37 °C. The plate was washed three times and then incubated with 100 µL HRP conjugate working solution for 30 min at 37 °C. Following five additional washes, 90 µL substrate reagent was added, and the reaction proceeded for 15 min (protected from light). The enzymatic reaction was stopped with 50 µL stop solution, and absorbance was immediately measured at 450 nm using a pre-warmed microplate reader.

Statistical analysis

When comparing data between two groups, if both groups follow a normal distribution and have equal variances, the t-test is used for statistical analysis; otherwise, the non-parametric rank sum test(e.g., Mann–Whitney U test or Wilcoxon rank-sum test) is employed. When comparing data across multiple groups, if the data meet the assumptions of normality and homogeneity of variance, analysis of variance (ANOVA) with Tukey’s or Dunnett post-hoc test for pairwise comparisons is used for statistical analysis. The Tukey test is employed when pairwise comparisons of mean differences among three or more groups are required. The Dunnett test is utilized when multiple treatment groups need to be compared with a single control group. If these conditions are not met, non-parametric tests (e.g., the Kruskal–Wallis H test) followed by Dunn’s post-hoc test for pairwise comparisons should be applied. (* < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001).