Human samples

Ethical approval for the study was granted by the Ethics Committee at the First Affiliated Hospital of Soochow University (Suzhou, China. No. 2023532), and informed consent was also obtained from all participants. Paraffin-embedded tissue specimens from UC or CD patients were collected from the First Affiliated Hospital of Soochow University. Comprehensive clinicopathological data for both UC and CD patients can be found in Additional file 1 Supplementary Table 1–2. Diagnosing CD or UC involves utilizing various methods such as clinical, biochemical, stool, endoscopic, imaging, and histological tests as recommended by European Crohn’s and Colitis Organization [ECCO] and European Society of Gastrointestinal and Abdominal Radiology [ESGAR] guidelines [2]. Ulcerative colitis disease activity and the Crohn's disease activity index (CDAI) were assessed according to established protocols [22].

Microarray data

Microarray data from GSE164985, GSE75214, GSE10616, GSE179285, GSE38713, GSE10191, GSE20881, GSE95095, GSE102133, and GSE186582 were obtained from the GEO database at http //www.ncbi.nih.gov/geo. We obtained the original data in the form of MINiML files. The analysis of gene expression was conducted utilizing the R software, specifically the ggplot2 package. The data extracted from the GEO database was normalized through log2 transformation and the microarray data was further normalized using the normalize quantiles function within the preprocessCore package in R software (version 3.4.1).

Mice and colitis models

The Ethics Committee of Soochow University (Suzhou, China) approved all animal experimental procedures (No.202311A0034). Male mice of the C57BL/6 or BALB/c strains, aged between 6 and 8 weeks, were acquired from the Laboratory Animal Center at Soochow University. A DSS-induced colitis model was established as indicated previously [23]. In brief, C57BL/6 mice were fed 2.0% DSS (#160,110, MP Biomedicals, USA) for 7 days. A TNBS-induced acute colitis model was established as previously described [23]. Following a 7-day period of skin sensitization with a 1% (wt/vol) TNBS presensitization solution, BALB/c mice were lightly anesthetized (intraperitoneal injection dosing of 80 µl/10 g body weight of ketamine/xylazine solution) and administered 2.5% (wt/vol) TNBS through a catheter inserted 4 cm into the colonic lumen. Control mice received a 100 μl dose of 50% ethanol.

In order to investigate the function of KLF4 in colitis, mice were treated with either KLF4 shRNA adenovirus (shKLF4-Adv, 1 × 1010 PFU/ml, 100 μl/mouse) or control adenovirus (control-Adv) via intracolonic administration right after receiving 2.0% DSS treatment. For TNBS-induced colitis model mice, on the day of TNBS presensitization, the first shKLF4-Adv (1 × 10^10 PFU/ml, 100 μl/mouse) or control-Adv enema was concurrently started.

In both models, mice were observed and weighed daily. Intestinal inflammation severity was evaluated with the disease activity index (DAI) according to previous descriptions [23]. The DSS-challenged mice were sacrificed at 7 days after DSS treatment, and the TNBS-treated mice were sacrificed at 3 days after 2.5% (wt/vol) TNBS treatment. The colons were removed for macroscopic examination, measuring the length of the colon, analyzing histopathology using hematoxylin–eosin (H&E) and alcian blue periodic acid–Schiff (AB-PAS) staining, and conducting cytokine analysis using RT‒qPCR.

Histology

Following treatment with 4% paraformaldehyde (#P1110, Solarbio, Beijing, China) for 24 h, the colon tissues of the mice were preserved in paraffin and cut into 5 µm slices. Following the guidelines provided by the manufacturer, H&E staining was performed using a staining kit from Beyotime in Shanghai, China. Histological scores for colitis induced by DSS and TNBS were calculated following the traditional protocol [23]. Furthermore, AB-PAS staining was carried out following the guidelines provided by the manufacturer of the product from Servicebio in Wuhan, China.

Cell culture and stimulation conditions

The colonic epithelial adenocarcinoma cell line Caco-2 was acquired from the American Tissue Culture Collection (ATCC, USA) and the colonic epithelial cell line NCM460 was purchased from Incell Corporation LLC (INCELL, USA). Caco-2 and NCM460 were grown in RPMI-1640 (Eallbio, Beijing, China) with 10% fetal bovine serum (FBS, Eallbio) and 1% penicillin‒streptomycin (Beyotime) in a humidified incubator with 5% CO2 at 37 °C. Intestinal epithelial monolayer barriers were constructed using Caco-2 cells. To investigate the role of KLF4 in intestinal epithelial barriers, Caco-2 monolayers were incubated with H2O2 (#108,597, Sigma-Aldrich, St. Louis, USA) for 4 h or TNF-α (#SRP3177, Sigma‒Aldrich)/IFN-γ (#SRP3058, Sigma‒Aldrich) mixtures for 18 h. To explore the synergistic impact of mesalazine (#HY-15027, MCE, New Jersey, USA) and KLF4 overexpression on the maintenance of intestinal epithelial barrier function, mesalazine at a concentration of 5 mmol/L and the KLF4 inducer APTO-253 (#HY-16291, MCE) at a 1 μmol/L concentration were added to Caco-2 cells and incubated for 24 h.

Cell transfection and infection

MiaoLingBio (Wuhan, China) designed and synthesized overexpression plasmids for KLF4, ALKBH5, and METTL3. GenePharma (Suzhou, China) provided commercial siRNAs targeting KLF4, ALKBH5, METTL3, YTHDF1/2/3, YTHDC1/2, and IGF2BP1/2/3 along with their corresponding controls. Caco-2 cells were grown in 6-well dishes and LipofectamineTM 3000 (#L3000001, Thermo Scientific, Waltham, USA) was utilized for transfecting Caco-2 cells with plasmids or siRNAs. GenePharma synthesized lentiviruses for KLF4 with both overexpression and knockdown capabilities, as well as their respective controls. Upon reaching 40% confluence, the Caco-2 cells were exposed to lentiviral particles with a multiplicity of infection (MOI) of 40. The effectiveness of the infection was assessed through quantitative real‑time polymerase chain reaction (qRT‒PCR) and Western blot analysis.

Immunohistochemistry (IHC)

Sections of human or mouse tissue embedded in paraffin were sliced to a thickness of 5 μm, followed by dewaxing, retrieval of antigens at high temperature, blocking of nonspecific antigens, and overnight exposure to primary antibodies at 4 °C, then incubated with secondary antibodies conjugated with HRP for 30 min at 25 °C. The antibodies utilized in the IHC assay are detailed in Supplementary Table 3. Detection was carried out using DAB substrate (#8059, CST, Danvers, USA).

Immunofluorescence (IF) staining

Paraffin-embedded sections of colon tissues were dehydrated in xylene and ethanol before being blocked with goat serum. For IF staining of Caco-2 cells, we cultured Caco-2 cells on glass coverslips the day before experimentation. The cells were treated with 4% paraformaldehyde and then exposed to 0.5% Triton X-100 for permeabilization. Afterward, the sections or cells were treated with primary antibodies overnight at 4 °C, then stained with secondary antibodies. Following staining of the nuclei with DAPI (#P0131, Beyotime), fluorescence images were captured using an immunofluorescence microscope. The antibodies utilized for immunofluorescence staining can be found in Supplementary Table 3.

Western blot

SDS lysis buffer (#P0013G, Beyotime) containing a protease and phosphatase inhibitor cocktail (#5872, CST) was used to lyse cells and tissues. The BCA protein assay kit (#P0011, Beyotime) was utilized to determine the protein concentration. The protein samples underwent SDS-PAGE and were then transferred onto PVDF membranes. After an overnight period of incubation with primary antibodies at 4 degrees Celsius, the membranes were rinsed and then treated with horseradish peroxidase (HRP)-linked secondary antibodies at 25 degrees Celsius for 2 h. The spots were observed with an improved chemiluminescence (ECL) setup.The antibodies employed in this investigation are detailed in Additional file 1 Supplementary Table 3.

qRT‒PCR

Total RNA was extracted from cells or tissues using TRIzol (#15,596,026, Thermo Scientific) according to the manufacturer’s instructions. Reverse transcription was carried out using the Transcript First Strand cDNA Synthesis Kit (#04897030001, Roche, Basel, Switzerland), and qPCR was performed using Fast Start Universal SYBR Green Master Mix (#12,239,264,001, Roche). Lithium chloride (8 M, 0.1 volume) was used to purified the samples from DSS-induced colitis model mice to prevent potential inhibition of qPCR by DSS [24]. mRNA expression was relatively quantified using the ΔΔCt technique. The control utilized was GAPDH. Three technical replicates were used in the PCR procedure. The primer sequences utilized are detailed in Additional file 1 Supplementary Table 4.

Transepithelial electrical resistance (TEER) measurement

For the TEER assay, a 24-well Transwell system (#725,101, NEST, Wuxi, China) was used. Caco-2 cells were placed in the top chamber with a concentration of 1 × 104 cells per 300 μl. Next, the lower chamber received 1.0 mL of medium. The TEER was measured on the specified day with an epithelial voltage resistance meter from Jingong Hongtai Technology Co., Ltd. in Beijing, China. TEER values were adjusted for background resistance stemming from the membrane insert and computed as Ω⋅cm2. Electrical resistance measurements were conducted until three consecutive readings were consistent.The formula used to calculate the TEER was (total resistance – blank resistance) × area.

Fluorescein isothiocyanate‑dextran permeability assay

In vivo and in vitro, intestinal permeability was evaluated using Fluorescein isothiocyanate-dextran (FD4; Sigma‒Aldrich) with a molecular weight of 4 kDa. For in vivo experiments, after forbidden food and water for 4 h, the mice were gavaged with 150 μl of FD4 solution (80 mg/mL solution). After a span of four hours, serum samples were obtained from the mice and analyzed for FD4 concentration with a TECAN Infinite F500 microplate reader configured with excitation/emission wavelengths of 485/535 nm. For the in vitro experiments, a 24-well Transwell system was used. Caco-2 cells were placed in the top chamber with a concentration of 1 × 104 cells per 300 μl. Next, the lower chamber received 1.0 mL of medium. After around 20 days of stability in TEER, Opti-MEM from Gibco in the USA was utilized to create a 2 mg/mL solution of FD4 powder. The upper transwell chamber was supplemented with FD4 solution, while the lower chamber received Opti-MEM. After incubating for 4 h, the TECAN Infinite F500 microplate reader was used to measure the fluorescence intensity emitted from the lower chamber at wavelengths of 485/535 nm.

TUNEL staining

Colon sections were subjected to TUNEL staining using a TUNEL kit (#PF00006, Proteintech, Wuhan, China) following the manufacturer's protocol. Cells positive for TUNEL staining were observed with a confocal microscope (FLUOVIEW FV3000, Olympus) and counted using ImageJ software.

Caspase3 activity assay

Caspase3 activity assays were performed using the Caspase3 Activity Assay Kit(#C1115, Beyotime). Briefly, a quantity of 10 mg of mouse colon tissue was combined with 100 μl of lysate, homogenized using a glass homogenizer on an ice bath, and subjected to lysis for a duration of 5 min. The resulting supernatants from the homogenate were collected through centrifugation at 16,000 g at 4 ºC, followed by determination of the protein concentration utilizing a Bradford protein assay kit. Subsequently, tissue lysates were incubated with Ac-DEVD-pNA (2 mM) at 37 °C for 2 h. After incubation, absorbance was read at 405 nm using a microplate reader. The enzymatic activity of caspase3 within the protein per unit weight of the sample is determined by dividing the quantity of pNA generated through catalysis in the sample by the protein concentration of the sample.

Chromatin immunoprecipitation (ChIP)

The ChIP experiment utilized a ChIP kit (Hyperactive pG-MNase CUT&RUN Assay Kit for PCR/qPCR (#HD101-01, Vazyme Biotech Co., Ltd., Nanjing, China) in accordance with the instructions provided by the manufacturer. 100,000 Caco-2 cells were gathered and exposed to ConA Beads Pro at ambient temperature. Following this, the cell-magnetic bead combination was left to incubate with primary antibodies for the entire night at a temperature of 4 °C. The pG-MNase enzyme mixture was made, flipped, and then chilled on ice. Next, the blend was introduced to the cell-magnetic bead compound and left to incubate while rotating for 1 h at 4 °C. Next, 100 μl of CaCl2 premix was introduced, followed by inverting the mixture multiple times to thoroughly blend the buffer and cell-magnetic bead complex. Next, 100 μl of stop buffer was introduced, followed by inversion of the mixture, placement in a 37 °C water bath, and an incubation period of 10–30 min. The stop buffer contained 10 pg of spike-in DNA fragments, which can be used for uniform calibration of experimental tests. FastPure gDNA Mini Columns were used for DNA extraction and purification. The qPCR utilized the purified DNA sequences for E-cadherin, Occludin, ZO-1, and Claudin-4 promoter regions as follows: E-cadherin Forward, 5ʹ-TAGAGGGTCACCGCGTCTAT-3ʹ Reverse, 5ʹ-TCACAGGTGCTTTGCAGTTC-3ʹ; Occludin Forward, 5ʹ-CCTGCTGGATGGCAACTAA-3ʹ Reverse, 5ʹ-AACGAAAGACTCCTGGGAAA-3ʹ; ZO-1 Forward, 5ʹ-TGTTTAGCAAAGCCGTCA-3ʹ Reverse, 5ʹ-CAATGCTGTCCCTCCAAC-3ʹ; and Claudin-4 Forward, 5ʹ- CTGGGGTGATGATGTCTCCAG-3ʹ Reverse, 5ʹ-CAGCTTCTCTGCTTGGCTAG-3ʹ. The fold change method was recommended for calculating the desired values, with the specific calculation procedure outlined as follows:

(1) The Ct values of the target gene and spike DNA in the treatment (anti-KLF4) group and negative control group (IgG) were determined.

(2) The ΔCt values for each group were determined by utilizing spike DNA as an internal control.

$$ \begin }\;}: \, \Delta }_}}} = }_}}} - }_}} \hfill \\ }\;}: \, \Delta }_}}} = }_}}} - }_}} \hfill \\ \end $$

(3) Taking negative control IgG as a reference, the 2−△△Ct values of each group were calculated.

$$ \begin } lative \, foldenrichment \, of\;anti - KLF4\left( \right) \, = \, 2^ \right)}} \hfill \\ } lative \, foldenrichment \, of\;IgG\left( \right) \, = \, 2^ \right)}} = 1 \hfill \\ \end $$

Apoptosis analysis

Apoptosis was assessed utilizing the 7-AAD/Annexin V Apoptosis Detection Kit I (#559,763, BD Bioscience, San Jose, USA). Caco-2 cells were plated in a 6-well plate at a concentration of 1 × 106 cells per well and grown until the formation of a monolayer representing the intestinal epithelial barrier. Subsequently, each cell sample was treated with 5 μl of PE-conjugated Annexin V and 5 μl of 7-AAD and incubated for 15 min at 4 °C in a light-free environment. Staining of cells was followed by flow cytometry (Beckman Coulter, California, USA). Each experimental procedure was conducted in triplicate. Annexin-V + /7-AAD- cells and Annexin-V + /7-AAD + cells were considered apoptotic cells. The data were analyzed using FlowJo statistical software (v10.62).

Determination of cellular reactive oxygen species (ROS) levels

Cellular levels of ROS were measured using a kit for detecting Reactive Oxygen Species (#S0033S, Beyotime). Caco-2 cell cultures were exposed to H2O2 (10 μmol/L) for a duration of 4 h in 6-well dishes. Cells were treated with 10 mmol/L DCFH-DA in medium without serum for 30 min at 37 °C. After washing with warm PBS, single-cell suspensions were obtained from the 6-well plates and analyzed via flow cytometry (Beckman Coulter). FlowJo statistical software (v10.62) was utilized for data analysis.

MeRIP assays

The MeRIP assay was conducted utilizing the EpiQuik™ CUT&RUN m6A RNA Enrichment (MeRIP) Kit (#P9018, Epigentek, Farmingdale, USA). Briefly, RNA fragments containing m6A were captured using a bead m6A capture antibody, followed by purification and elution of the enriched RNA. Subsequently, RT‒qPCR was used to quantify changes in the methylation of target genes, KLF4 m6A motif locations are shown in Additional File 2–2.

RNA immunoprecipitation (RIP) assay

The Magna RIP kit (#17–700, Sigma‒Aldrich) was utilized to conduct the RIP assay following the instructions provided by the manufacturer. In brief, cellular cleavage was performed on ice using RIP lysis buffer, followed by overnight incubation of whole-cell lysates with magnetic protein A/G beads and either 5 µg of normal rabbit IgG or primary antibodies (anti-METTL3, anti-METTL14, anti-WTAP, anti-FTO, anti-ALKBH5 and anti-YTHDF2) at 4 °C. Before the purification step of RNA, 50 μL of the bead suspension was taken during the last wash to test the efficiency of immunoprecipitation by Western blotting. Afterwards, the RNA–protein compound underwent ongoing eluent washing and was exposed to proteinase K at 55 °C for a duration of 30 min. The bound RNA was then extracted for qRT‒PCR analysis. The primer sequences are provided in Supplementary Table 4.

RNA stability assay

Caco-2 cells were placed in a 6-well dish and incubated for a day. Afterward, the cells were exposed to actinomycin D (10 µg/ml, #A9415, Sigma‒Aldrich) for durations of 1, 2, 3, or 4 h. RNA was isolated from the cells in preparation for future qRT‒PCR tests.

Enzyme‑linked immunosorbent assays (ELISA)

The concentration of IL-1β and IL-6 was measured using commercially-available ELISA Kits (ExCell Bio, Shanghai, China). Briefly, Caco-2 cells were cultured in a 6-well plate until a monolayer resembling the intestinal epithelial barrier was formed. After the cells were treated with TNF-α/IFN-γ mixtures for 24 h, the cell supernatant was collected via the centrifugation. The concentration of IL-1β and IL-6 in the cell supernatant was measured using the corresponding ELISA Kit following the manufacturer’s instructions.

Dual-luciferase reporter assays

A renilla luciferase plasmid and a pGL3-Basic-Luciferase reporter vector containing the promoter sequence of E-cadherin (pGL3-E-cadherin) were obtained from MiaoLingBio. KLF4 knockdown or overexpression caco-2 cells were cultured in a 24-well plate and were co-transfected with the pGL3-E-cadherin vectors (250 ng/well) and Renilla luciferase plasmid (10 ng/well). After 48 h post-transfection, the cells were lysed using passive lysis buffer, and the luciferase activity was quantified using the Dual-Luciferase Reporter Assay System kit (#E1910, Promega, Madison, USA). The Renilla luciferase activity wasserved as an internal reference.

Statistical analysis

GraphPad Prism 9 (La Jolla, CA, USA) was used to analyze all the statistical data in this study. The data are reported as the mean ± SD. Statistical analyses were conducted using unpaired two-tailed Student’s t tests, one-way ANOVA, and nonparametric Mann‒Whitney U tests. Spearman’s correlation was utilized for nonparametric correlation analysis. Significance was determined at p < 0.05. The data presented are representative of a minimum of three independent experiments.