Helicobacter pylori CagA elevates FTO to induce gastric cancer progression via a "hit-and-run" paradigm.

1
BACKGROUND
Gastric cancer (GC) ranks as the fifth most prevalent malignant tumor and is the fourth leading cause of cancer‐related death [1, 2]. Chronic infection with Helicobacter pylori (H. pylori) is a significant contributor to the development of GC [3, 4]. The pathohistological outcome of H. pylori infection is influenced by both the bacterium's virulence factors and the host's genetic characteristics [5, 6]. Among these virulence factors, cytotoxin‐associated gene A (CagA) stands out as the primary factor, which can be delivered into host cells via the type IV secretion system (T4SS), thereby triggering a cascade of oncogenic signaling pathways [7, 8, 9, 10]. Strains of H. pylori that harbor the cagA gene (cagA
+) are associated with more pronounced gastric mucosal damage and exhibit a higher propensity for inducing GC compared to cagA‐negative strains [6, 7, 11].
Given the prominent role of H. pylori in the progression of GC, strategies for large‐scale screening and eradication of H. pylori have been implemented with the aim of preventing and treating GC [1, 12]. A wealth of prospective studies and meta‐analyses have demonstrated that the eradication of H. pylori is associated with a reduced risk of GC [12, 13, 14, 15]. Nevertheless, pathohistological findings suggest that successful eradication of H. pylori failed to prevent the development of GC [16, 17]. Infection with cagA
+
H. pylori can elicit genomic and epigenetic alterations in host cells that may perpetuate tumor progression even in the absence of the pathogen [18]. Consequently, CagA is hypothesized to mediate carcinogenesis through a “hit‐and‐run” mechanism [18, 19], although the precise mechanism underlying this process is not yet fully understood.
N6‐methyladenosine (m6A) modification is a dynamic process that modulates gene expression in response to environmental stimuli [20, 21, 22] and plays an essential role in the development of tumors [23, 24, 25, 26]. Hence, we proposed the hypothesis that infection with cagA
+
H. pylori induces irreversible alterations in host cell m6A methylation, thereby facilitating GC progression in a “hit‐and‐run” manner. To validate this hypothesis, we conducted a series of investigations to delineate the role and underlying molecular mechanisms of m6A methylation in cagA
+
H. pylori infection. Furthermore, we assessed the therapeutic efficacy of specific m6A small‐molecule inhibitors in impeding GC progression.

2
MATERIALS AND METHODS
2.1
Cell and H. pylori culture
MKN45 and AGS cell lines were obtained from the Japanese Collection of Research Bioresources (Osaka, Japan) and the American Type Culture Collection (Manassas, VA, USA), respectively. The cells were cultivated in Dulbecco's Modified Eagle Medium (#C11965500BT, Gibco, Burlington, MA, USA) supplemented with 10% fetal bovine serum (FBS, #A5669701, Gibco). All of the cell lines were confirmed by genotyping and were tested regularly for mycoplasma contamination.
The H. pylori wild‐type (WT) cagA
+ strain (National Collection of Type Cultures [NCTC] 11637, cagA
+
H. pylori) and H. pylori isogenic mutant strain (NCTC 11637ΔcagA, ΔcagA H. pylori) were provided by Professor Quanming Zou and Yuan Zhuang, who come from Army Military Medical University (Chongqing, P. R. China). The strains were inoculated onto H. pylori culture plates enriched with 5% sheep blood (#HB8646, HOPEBIO, Qingdao, Shandong, P. R. China) and incubated at 37°C under microaerobic conditions (10% CO2, 5% O2, 85% N2), utilizing a microaerophilic system (Becton, Dickinson and Company, San Diego, CA, USA).

2.2
Clinical tissue samples and organoid culture
Cancerous tissues were collected from fifteen patients with pathologically confirmed GC. The utilization of patient samples was approved by the Medical Ethics Committee of Xinqiao Hospital of Army Medical University (Approved No. 2024‐083‐01). Using an H. pylori antibody classification assay kit (#CP01, Blot, Shenzhen, Guangdong, P. R. China), we systematically categorized all patient samples into CagA‐positive (n = 8) or CagA‐negative (n = 7) groups. This classification was subsequently verified through immunohistochemical (IHC) detection of the CagA protein using CagA antibody (#sc‐28368, Santa Cruz Biotech, Dallas, TX, USA). We also obtained RNA‐seq data from GC patients in The Cancer Genome Atlas (TCGA) database, with the dataset project ID TCGA‐STAD.
Fresh GC tissues were washed in phosphate‐buffered saline (PBS) with 1% penicillin‐streptomycin and cut into 1‐3 mm2 fragments. A digestion mixture was prepared with 10 mL of organoid culture medium (#K2179‐GC, BioGenous, Suzhou, Zhejiang, P. R. China), 1 mg of collagenase XI (#C7657, Sigma Aldrich, Burlington, MA, USA) and 10 mg of dispase II (#D4693, Sigma Aldrich). The fragments underwent enzymatic digestion for about 1 hour. Gentle pipetting dispersed the fragments in PBS. Fragments were collected and resuspended in Matrigel (#356231, Corning, Bedford, NY, USA) at 50 µL per well, seeded into a prewarmed 24‐well plate, and allowed to solidify at 37°C for 10 minutes. After solidification, 500 µL of complete organoid culture medium was added to each well. The medium was refreshed every two to three days. Three CagA‐negative GC organoids were successfully cultured for further experiments.

2.3
Co‐culture of H. pylori with cells and organoids
GC cells at the logarithmic growth stage were seeded into 6‐well plates and cultured in an antibiotic‐free medium. H. pylori was then harvested and introduced into the GC cells at a multiplicity of infection (MOI) of 100:1. After 24 hours of co‐culture, the GC cells were collected for further analysis.
We performed co‐cultures of H. pylori with organoids as described in previously published literature [27]. Briefly, well‐formed organoids were selected and centrifuged at 300 ×g for 5 minutes. The organoids were then resuspended in a culture medium and transferred to 15 mL tubes. H. pylori was scraped with an inoculation loop and suspended in 1 mL of sterile PBS. After centrifuging at 5000 rpm for 5 minutes, the bacteria were resuspended in 1 mL of sterile PBS. This suspension was then mixed with the organoids in a 15 mL tube and incubated for 3 hours. Thereafter, the organoids were resuspended in Matrigel and placed into individual wells of a 24‐well plate. Each well was then filled with 500 µL of standard organoid culture medium.

2.4
Tissue microarray (TMA) and IHC
Human GC TMAs with documented H. pylori infection status were purchased from Shanghai Outdo Biotech Co. Ltd. (#HStmA180Su30, Shanghai, P. R. China). This company has provided ethical approval for these TMAs (Approved No. YBM‐05‐02). The TMA includes cancerous tissues and their corresponding para‐carcinoma counterparts from a cohort of 83 GC patients, supplemented by cancer tissues from 14 additional patients that lack matched para‐carcinoma tissues. IHC was performed using a commercial IHC staining kit (# IHC001 and IHC003, Boster, Pleasanton, CA, USA), IHC staining strictly following the instructions. In brief, the TMAs were deparaffinized with xylene and antigenically repaired with sodium citrate and serum blocking. The arrays were then probed with primary antibodies specific for either fat mass and obesity‐associated protein (FTO, Abcam, Cambridge, MA, USA; 1:100) or CagA (Santa Cruz Biotech, 1:100), with incubation overnight. Subsequently, the TMAs were exposed to the corresponding secondary antibodies for 1 hour. The 3,3'‐Diaminobenzidine (DAB) was used for color development, and the resulting slides were photographed using an ortho‐microscope (Nikon, Tokyo, Japan). The staining intensity was quantified by measuring the integrated optical density of the slides with the aid of ImageJ software (National Institutes of Health, Bethesda, MD, USA).

2.5
Animal models
All of the animal studies were conducted in compliance with protocols approved by the Laboratory Animal Welfare and Ethics Committee of Army Military Medical University (Approval No. AMUWEC20235107). Four‐week‐old male nude mice (#401, Vital River, Beijing, P. R. China) were housed in a Specific Pathogen Free (SPF) laboratory environment and randomly assigned to groups of shNC, shFTO, NC, FTO‐WT‐oe, FTO‐mut1‐oe, and FTO‐mut2‐oe. Following induction of anesthesia, the mice were intravenously inoculated with 5 × 106 GC cells via the tail vein. Bioluminescent imaging was performed four weeks post‐inoculation to assess lung metastasis. D‐luciferin potassium salt (#P1042, 150 mg/kg, Promega, Madison, WI, USA) was administered intraperitoneally, and after a 10‐minute interval, bioluminescent signals were captured using Bruker Molecular Imaging Software (Bruker, Billerica, MA, USA). At the end of the study, all animals were humanely euthanized. Hematoxylin‐eosin (HE) staining was performed to confirm the lung metastatic burden.
For in vivo therapeutic experiments, CagA‐tet‐on or negative control (NC‐tet‐on) GC cells were administered via the tail vein injection. The mice were orally administered doxycycline (DOX, 500 mg/kg once daily; #HY‐N0565, MedChemExpress, Monmouth Junction, NJ, USA) and intraperitoneally injected with MA (50 mg/kg, once every 3 days, #HY‐B1320, MedChemExpress) for 30 days. Bioluminescent imaging and HE staining were carried out at the end of the treatment regimen.

2.6
Dot blotting
Total RNA was extracted from GC cells or tissues, and prepared RNA dilutions were subjected to a denaturation step at 95°C for three minutes. A Hybond‐N+ membrane (#RPN303B, Millipore, Billerica, MA, USA) was utilized for the immobilization of 2 µL of RNA. The RNA on the membrane was crosslinked using an ultraviolet crosslinker (Stratagene, Santa Clara, CA, USA) at an intensity of 1200 µJ for two rounds, each lasting about 25 to 50 seconds. Post crosslinking, the membranes were incubated with a 1% solution of bovine serum albumin (BSA) for 1 hour. m6A antibody (#ab284130, Abcam, 1:1000) and appropriate secondary antibodies were subsequently applied to the membranes at recommended dilution ratios. The enhanced chemiluminescence reagent (ECL, #P0018S, Beyotime, Shanghai, P. R. China) was prepared for luminescence detection. Membranes were photographed with the help of a chemiluminescence system (Bio‐Rad, Hercules, CA, USA). To assess the total RNA content as a loading control, membranes were soaked for two hours in a 0.02% methylene blue solution (#S0296, Beyotime). Following this, membranes were rinsed with enzyme‐free water, and images were captured with a camera.

2.7
Organoid immunofluorescence and HE staining
Following the disposal of the media, organoids were washed twice with PBS. We used 4% paraformaldehyde to fix these organoids for 20 minutes at room temperature. After being softly trembling, the organoids were transferred to 1.5 mL centrifuge tubes. After which, samples were incubated with Triton‐X 100 at 4°C for 30 minutes, treated with 1% BSA for 1 hour, and incubated with diluted primary and secondary antibodies at 4°C according to the manufacturer's directions. The antibodies used in this investigation are listed in Supplementary Table S1. Following a 10‐minute at room temperature incubation period with 4',6‐diamidino‐2‐phenylindole (DAPI), the organoids were washed with PBS. The organoids were then resuspended in 20 to50 µL of anti‐fluorescence quenching sealing agent and examined under a microscope.
Characterization of organoids was achieved through HE and periodic acid‐schiff (PAS) staining. For HE, the organoids were initially fixed in a 4% paraformaldehyde solution, followed by embedding in paraffin for sectioning. Sections were dehydrated with alcohol, stained with hematoxylin, rinsed, and treated with a bluing agent to enhance nuclear contrast. Eosin was then applied to stain cytoplasm and connective tissue. The slides were dehydrated, cleared, and mounted with neutral balsam for microscopic analysis. In the PAS process, the sections were oxidized with periodic acid to reveal glycogen. After rinsing away excess oxidizer, Schiff's reagent was applied for color development. The reaction was stopped with a wash, followed by counterstaining with hematoxylin to delineate cell nuclei. Finally, the slides were dehydrated, cleared, and mounted with neutral balsam for examination under a microscope.

2.8
Plasmid construction
Lentiviral constructs overexpressing wild‐type FTO and mutant FTO were generated as previously described [28]. Knockdown lentiviral plasmids, such as shFTO and sh heparin‐binding EGF‐like growth factor (HBEGF), were cloned and integrated into the pPurGreen‐shRNA lentiviral vector (#SI505A‐1, System Biosciences, San Francisco, CA, USA) using the BamHI (# D6053, Beyotime) and EcoRI (#D6239, Beyotime) restriction enzymes. The target sequences for all shRNAs were detailed in Supplementary Table S2. The open reading frames (ORFs) of Flag‐tagged Jun proto‐oncogene (JUN), HBEGF, and HBEGF‐mut1/2 were amplified using the primers listed in Supplementary Table S3, with Flag tags incorporated during PCR and subsequently cloned into pCDH lentiviral vectors (#CD510B‐1, System Biosciences) via EcoRI (#D6329, Beyotime) and NotI (#D6497, Beyotime) enzymatic sites. The coding sequence of the cagA gene from cagA
+
H. pylori RNA underwent reverse transcription, amplification, cloning, and was then inserted into tet‐on lentiviral vectors (#030351, Clontech, Palo Alto, CA, USA), utilizing BamHI and EcoRI for the construction of CagA‐tet‐on lentiviral plasmids. All plasmid constructs were verified by DNA sequencing to ensure accuracy. The authenticated plasmids were co‐transfected with packaging vectors psPAX2 (#12260, Addgene, Watertown, MA, USA) and pMD2.G (#12259, Addgene) into HEK293T cells to produce lentiviral particles, which were collected at 48 and 72 hours post‐transfection. For establishing stable cell lines, GC cell lines MKN45 and AGS were transfected with the harvested virus in the presence of 4 µg/mL polybrene to enhance transduction efficiency.

2.9
Western blotting
The cells and GC tissues were repeatedly washed with PBS, and protein extraction was obtained using radioimmunoprecipitation assay (RIPA) buffer (#P0013B, Beyotime). Protein concentration was quantified using a BCA Protein Assay Kit (#P0010, Beyotime). After sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE) separation, the proteins were transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore). The membranes were then incubated with a blocking solution containing 5% BSA (#23210, Sangon, Shanghai, P. R. China) for 1 hour at room temperature. Following the step, membranes were incubated with primary antibodies, diluted according to the recommended ratios, and then with horseradish peroxidase (HRP)‐conjugated secondary antibodies, as listed in Supplementary Table S1. After this, the membrane was uniformly treated with an appropriate amount of ECL reagent (#P0018S, Beyotime). Protein bands were then visualized with a chemiluminescence system (Bio‐Rad). The intensity of the bands was quantitatively evaluated using ImageJ software.

2.10
Quantitative real‐time PCR (qRT‐PCR)
RNAiso reagent (#9109, Takara, Kusatsu, Japan) was used for the extraction of total RNA from both tissue and cells. A PrimeScript™ RT Reagent Kit with gDNA Eraser (# RR047A, Takara) was used for reverse transcription. To evaluate RNA expression levels, qPCR was conducted using the SYBR® Premix Ex Taq Kit (# RR820A, Takara) according to the manufacturer's instructions. The relative expression of the target genes was calculated via the 2−△△CT method, using glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) as the housekeeping gene [29]. Supplementary Table S4 lists the sequences of the primers used.

2.11
RNA sequencing (RNA‐Seq), Methylated RNA immunoprecipitation sequencing (MeRIP‐seq) and MeRIP‐qPCR
MeRIP‐seq experiments were performed according to the procedure described in the literature [26]. Briefly, total RNA was extracted and purified, followed by denaturation and fragmentation into approximately 100‐nt fragments. These RNA fragments were then incubated with protein A/G magnetic beads (#78609, Thermo Scientific, Waltham, MA, USA) in conjunction with an anti‐m6A antibody (#ab151230, Abcam). The input RNA and methylated RNA fragments, selectively bound to the m6A antibody, were eluted and prepared for subsequent sequencing analysis (Unionchuan, Hangzhou, Zhejiang, P. R. China). Genes with a log2 Fold Change (FC) > 1 and P < 0.01 were identified as differentially expressed genes (DEGs). Genes with P < 0.05 were identified as m6A‐marked genes. Gene Ontology (GO) classification enrichment of DEGs was conducted using the Sangerbox database (http://sangerbox.com). GSVA analysis was performed via the Gene Set Cancer Analysis (GSCA) website (http://bioinfo.life.hust.edu.cn/GSCA/#/).
The specific m6A sites in the mRNA of HBEGF were predicted using a sequence‐based N6‐methyladenosine (m6A) modification site predictor, SRAMP (http://www.cuilab.cn/sramp). MeRIP‐qPCR was performed to verify these predicted sites, employing a similar methodology to that of MeRIP‐seq with certain procedural adaptations. Specifically, total RNA was fragmented into 200‐nt pieces. Primers for PCR were custom‐designed to target candidate sites of interest (Supplementary Table S5). The enrichment of RNA fragments bound to the m6A antibody in each group was quantified via qRT‐PCR. The input RNA served as an internal control for normalization.

2.12
RNA immunoprecipitation (RIP)
A Magna RIP™ RNA Binding Protein Immunoprecipitation Kit (#17‐700, Millipore) was used for the experiment. Briefly, the cell lysates were incubated at 4°C overnight with A/G magnetic beads (Millipore) in the presence of either an IgG control antibody or a YTH N(6)‐methyladenosine RNA binding protein 2 (YTHDF2) (#24744‐1‐AP, Proteintech, Chicago, IL, USA) antibody. After being thoroughly washed, the A/G beads were harvested via a magnetic rack. RNA was then isolated, and qRT‐PCR was performed to quantify the enrichment of HBEGF RNA associated with YTHDF2 or IgG antibodies.

2.13
Identification of transcription factors (TFs) potentially interacting with FTO
The Animal TFDB (http://bioinfo.life.hust.edu.cn/AnimalTFDB/) and ChIP‐Atlas (https://chip‐atlas.org/) databases were used to identify TFs that may interact with the FTO promoter region. Further analysis of the correlation between the expression of these candidate TFs and FTO in our RNA‐seq data from GC cells post‐cagA
+
H. pylori infection was conducted. Additionally, the expression correlation of the candidate TFs and FTO in the data derived from Gene Expression Omnibus (GEO, accession ID: GSE62254, GSE26942, GSE15459, GSE84437) and TCGA‐STAD cohort was assessed. TFs that can act on the FTO promoter and were associated with FTO expression were selected for further assessment. The JASPAR (https://jaspar.genereg.net/) and Animal TFDB databases were selected to further predict the binding sites of the screened TFs with the FTO promoter. Based on the predicted results, FTO promoter‐specific primers were designed for subsequent analysis.

2.14
Chromatin immunoprecipitation (ChIP) assay
A Pierce Magnetic ChIP Kit (#26157, Thermo Scientific) was used for examinations according to the instructions. Initially, cells from each group were fixed with a 1% formaldehyde solution, followed by crosslinking and sonication. A portion of the lysate, 10 µL, was reserved as an input control and stored at ‐20°C. The remaining lysate was then incubated with either a JUN‐specific antibody or IgG antibody overnight at 4°C, each at a final concentration of 10 µg/mL. Subsequently, 20 µL of A/G beads was added to the mixture, which was gently agitated for 2 hours at 4°C. The protein‐chromatin complexes were carefully eluted, and the associated DNA was extracted after thorough washing. The immunoprecipitated DNA was subsequently analyzed using FTO promoter‐specific primers (Supplementary Table S6) to amplify the region of interest via PCR.

2.15
Enzyme‐linked immunosorbent assay (ELISA)
The concentration of HBEGF was quantified using an HBEGF ELISA Kit (#EK14428, Signalway Antibody, College Park, MD, USA) according to the manufacturer's instructions. In brief, 100 µL of diluted standards and samples were injected into corresponding plate wells. The plate was then sealed with paper and incubated at 37°C for 2 hours. After disposal of the liquid, 100 µL of detection reagent A's working solution was dispensed into each well. The plate was resealed and incubated for 1 hour at 37°C. Post incubation, the solution was removed, and each well was rinsed with 300 µL of the provided wash solution. This washing process was repeated a total of three times. Subsequently, 100 µL of detection reagent B's working solution was added to each well and incubated for 1 hour. Next, 90 µL of the substrate mixture was introduced to the wells and incubated for 15 to 25 minutes. The reaction was quenched by the immediate addition of 50 µL of stop solution. Finally, the quantitative assessment was conducted according to the absorbance at 450 nm with a microplate reader (Thermo Scientific).

2.16
Migration and invasion assays
Assays for migration were performed utilizing 24‐well Transwell plates with 8‐µm pores (#3428, Corning). The upper layer of each chamber was loaded with 5 × 104 cells in the logarithmic phase of growth. The lower compartment was filled with 500 µL of complete growth media containing 10% FBS. The plates were incubated for 24 hours at 37°C in a 5% CO2 atmosphere to allow cell migration. After being fixed with a 4% paraformaldehyde solution and stained with 1% crystal violet, the lower surface of the membrane was examined using a light microscope to quantify the number of migrated cells. For the invasion assay, precooled Matrigel (#354234, Corning) was equally applied to the upper layer of a chamber in 24‐well Transwell plates. The instructions for the migration assay were followed for subsequent procedures.

2.17
RNA stability assay
GC cells were plated in six‐well plates and allowed to adhere for 24 hours, at which point this was designated as the 0‐hour time point. Subsequently, one well was harvested, and 5 µg/mL actinomycin D (#A9415, Sigma Aldrich) was introduced to the remaining wells. At 2‐, 4‐, and 6‐hour post‐treatment, the cells from each time point were collected. Total RNA was extracted with RNAiso (#9109, Takara), and qRT‐PCR was used to determine the quantity of remaining RNA.

2.18
Dual‐luciferase reporter assay
MKN45 cells were transiently transfected with FTO reporter (Sangon) in 24‐well plates utilizing Lipo8000™ (#C0533, Beyotime). After a 4‐hour post‐transfection incubation period, the cells were infected with H. pylori at a MOI of 100:1 for 24 hours. Subsequently, the cells were washed with PBS and lysed for a dual‐luciferase assay (#E1960, Promega, Madison, MI, USA). The reporter activities were quantified using a microplate reader (Thermo Scientific). For experiments involving CagA and JUN, MKN45 cells with cagA overexpression or JUN knockdown and their respective controls were also transfected with FTO reporter. The dual‐luciferase activity assay was performed 24 hours later.

2.19
RNA pull‐down and mass spectrometry analysis
The biotin‐labeled probe of HBEGF (GenePharma, Shanghai, P. R. China) was diluted to a concentration of 20 pmol. This probe was then mixed with the cell lysate and streptomycin‐coated magnetic beads (#65001, Thermo Scientific). The mixture was incubated at 4°C overnight. After sufficient washing, the RNA and associated proteins were eluted from the beads. These eluted samples were then prepared for further analysis via immunoblotting and mass spectrometry.

2.20
Bioinformatic analysis
RNA‐seq data from 408 GC patients within the TCGA cohort (the dataset project ID: TCGA‐STAD) and detailed clinical features and prognostic information of 388 patients were downloaded from the UCSC Xena database (http://xenabrowser.net) on October 21, 2022. Patients were categorized into high and low expression groups based on the median gene expression value. Overall survival (OS) from the date of diagnosis was evaluated using Kaplan‐Meier curves and assessed with the log‐rank test. Expression profiles from the GEO datasets (accession ID: GSE62254, GSE29272, GSE14210, https://www.ncbi.nlm.nih.gov/geo/) were selected to validate relevant results. High and low expression groups were determined based on the optimal cutoff values via Kaplan‐Meier Plotter (http://kmplot.com/analysis/), and survival curves were subsequently generated. For gene set enrichment analysis (GSEA) analysis, GSEA software version 3.0 was obtained from the official website (http://software.broadinstitute.org/gsea/index.jsp). RNA‐seq data from FTO knockdown GC cells, retrieved from the GEO dataset (accession ID: GSE178697), were selected for analysis. The c2.cp.kegg.v7.4.symbols.gmt gene set collection was downloaded from the Molecular Signatures Database (http://www.gsea‐msigdb.org/gsea/downloads.jsp) to evaluate the enrichment of genes within relevant pathways.
The (epithelial‐mesenchymal transition) EMT score was calculated as: EMT score = (Fn1 + Vim + Zeb1 + Zeb2 + Twist1 + Twist2 + Snai1 + Snai2 + Cdh2) ‐ (Cldn4 + Cldn7 + Tjp3 + Muc1 + Cdh1) [35]. The correlation between FTO expression and EMT scores in GC tissues from the TCGA‐STAD cohort was assessed. EMT‐associated genes were downloaded from the EMTome database (http://www.emtome.org/). The correlation between HBEGF levels and the EMT score in the majority of tumor types was assessed via the EMTome database.

2.21
Statistical analysis
GraphPad Prism (version 9.0) was utilized for statistical tests. The results were displayed as average values and standard deviations (SD). The significance of the variance between the two groups was evaluated using either paired or unpaired Student's t‐test. The variances between multiple groups were evaluated using either one‐way or two‐way ANOVA. Pearson correlation tests were employed to assess the interrelations between FTO, HBEGF, and YTHDF2 levels. Survival for every group of GC patients was determined through Kaplan‐Meier curves and analyzed using the log‐rank test. A P value below 0.05 was acknowledged as statistically significant.

3
RESULTS
3.1
cagA
+
H. pylori mediated m6A modification and induced FTO expression
To determine the impact of cagA
+
H. pylori infection on m6A modification, we conducted an analysis on global m6A levels in GC tissues. Our results demonstrated that CagA‐positive GC samples exhibited significantly reduced global m6A levels in mRNA compared to those CagA‐negative samples (Figure 1A, Supplementary Figure S1A‐C). This finding was further confirmed in vitro using GC cells. Cells infected with the wild‐type (WT) H. pylori cagA
+ strain (cagA
+
H. pylori) displayed lower m6A levels when compared with both uninfected cells and cells infected with H. pylori isogenic mutant strains lacking cagA (ΔcagA H. pylori) (Figure 1B‐C).
m6A modifications were subject to dynamic regulation by a suite of m6A regulators [30, 31], with FTO exhibiting the most pronounced up‐regulation in response to cagA overexpression (Supplementary Figure S1D). This trend was corroborated in our analysis of GC tissues, where FTO expression was significantly higher in CagA‐positive GC tissues when compared to CagA‐negative GC tissues (Figure 1D, Supplementary Figure S1E). Further in vitro studies revealed that both mRNA and protein levels of FTO were significantly higher in cagA
+
H. pylori‐infected GC cells compared to the uninfected group (Mock) and the ΔcagA H. pylori‐infected group (Figure 1E‐F). Consistent with these findings, FTO levels were significantly greater in H. pylori‐infected GC tissues in the TCGA‐STAD cohort compared to H. pylori‐uninfected GC tissues (Supplementary Figure S1F). To simulate the infection of cagA
+
H. pylori in the host environment, human GC organoids were cultured and characterized using HE and PAS staining (Supplementary Figure S1G). The successful construction of GC organoids was verified by the expression of common GC markers, including carcinoembryonic antigen (CEA), cytokeratin 20 (CK20), epithelial cell adhesion molecule (EpCAM) and cytokeratin 7 (CK7), through immunofluorescence (Supplementary Figure S1H). Organoids derived from CagA‐negative GC tissues were selected for inoculation with cagA
+
H. pylori. Notably, infection with cagA
+
H. pylori led to a significant increase in FTO expression in GC organoids compared to the Mock and ΔcagA H. pylori infection groups (Figure 1G).
The expression of FTO in the TCGA‐STAD cohort was conducted to explore its clinical implications in GC. Compared to normal gastric tissues, GC tissues presented a markedly higher increase in FTO expression (Figure 1H). Robust correlations were identified between elevated FTO mRNA levels and key clinicopathological features, including lymph node metastasis, distant metastasis, and tumor stage (Figure 1I). In addition, when compared with the other subtypes of GC, the genomically stable subtype, which is associated with a more aggressive disease phenotype [34], exhibited significantly elevated FTO expression levels (Supplementary Figure S1I). The OS of GC patients with high FTO expression was significantly lower than those of patients with low FTO expression in both the TCGA‐STAD cohort and GEO datasets (Figure 1J, Supplementary Figure S1J). These observations collectively suggest that the upregulation of FTO expression, potentially mediated by cagA
+
H. pylori infection, may play a pivotal role in the accelerated progression of GC.

3.2
Elevated FTO was necessary for CagA‐induced GC metastasis
To explore the biological functions of CagA‐induced FTO upregulation in GC progression, we retrieved RNA‐seq data of FTO knockdown GC cells from the GEO dataset (GSE178697) [32]. KEGG enrichment analysis revealed that the knockdown of FTO led to significant alterations in gene expression profiles, with a pronounced enrichment of genes implicated in EMT‐related processes, such as cell adhesion, cell polarity, and ECM‐receptor interaction (Figure 2A). These findings hint at a potential role of FTO in the induction of EMT in GC cells, thereby possibly facilitating their metastatic ability. Analysis of the GC dataset GSE62254 revealed that FTO expression levels were significantly elevated in tissues from patients exhibiting distant metastasis (M1) compared with those without metastasis (M0) (Figure 2B). Furthermore, the metastatic and invasive capabilities of GC cells were considerably attenuated following FTO knockdown (Figure 2C‐D, Supplementary Figure S2A‐B). To verify the demethylase function of FTO, we constructed lentiviruses harboring wild‐type (WT) and mutant (FTO‐mut1/FTO‐mut2) FTO sequences (Supplementary Figure S2C‐D), as previously described [28], and established stable cell lines with these lentiviruses (Figure 2E, Supplementary Figure S2E). Our results indicated that, in contrast to the mutants, FTO‐WT‐oe significantly augmented GC cell metastasis and invasion (Figure 2F, Supplementary Figure S2F).
Additionally, we observed that the migratory and invasive capacities of MKN45 and AGS cells were substantially increased upon infection with cagA
+
H. pylori (Figure 2G, Supplementary Figure S2G). Importantly, the metastatic effects induced by cagA
+
H. pylori infection were mitigated by FTO knockdown (Figure 2G, Supplementary Figure S2G), suggesting that the enhancement of GC cell migration by cagA
+
H. pylori is FTO‐dependent.

3.3
CagA transcriptionally promoted FTO expression through JUN
We have previously demonstrated that CagA was capable of elevating both mRNA and protein levels of FTO (Figure 1E‐F). Furthermore, cagA
+
H. pylori infection did not influence mRNA stability (Figure 3A) but markedly enhanced FTO promoter activity (Figure 3B‐C). This prompted us to utilize the Animal TFDB and ChIP‐atlas databases to identify the TFs that may interact with the FTO promoter region. Through further analysis of the correlation between the expression of these candidate TFs and FTO in our RNA‐seq data from GC cells post‐cagA
+
H. pylori infection, along with GC datasets from GEO and TCGA‐STAD, we identified seven TFs as potential regulators of FTO expression (Figure 3D). Notably, among these, knockdown of JUN resulted in a significantly reduced expression of FTO (Figure 3E, Supplementary Figure S3A). We proceeded to investigate nine putative JUN binding sites within the FTO promoter, identified through JASPAR and Animal TFDB, and designed specific primers for validation (P1‐P9) (Figure 3F). ChIP‐PCR confirmed that the P3, P7, and P8 fragments were significantly enriched upon cagA
+
H. pylori infection (Figure 3G) and JUN overexpression (Figure 3H), indicating these regions as JUN‐specific binding sites in the FTO promoter.

cagA
+
H. pylori infection upregulates JUN expression via MAPK, and the use of MAPK1/3 inhibitors attenuated the nuclear translocation of JUN [33]. Interestingly, genes with altered expression following cagA
+
H. pylori infection, as identified from our RNA‐Seq data, were found to be significantly enriched in the mitogen‐activated protein kinase (MAPK) pathway (Figure 3I). Kaplan‐Meier analysis of the GSE62254 dataset revealed a significant positive relationship between elevated JUN levels and poor OS in GC patients (Supplementary Figure S3B). We further confirmed that cagA
+
H. pylori infection promotes JUN expression (Supplementary Figure S3C). JUN knockdown resulted in decreased FTO promoter activity (Figure 3J) and reduced mRNA level of FTO (Figure 3K, Supplementary Figure S3D), while overexpression of JUN promoted FTO expression (Figure 3L). Additionally, a positive correlation was observed between FTO and JUN expression levels (Figure 3M, Supplementary Figure S3E‐G). The increase in FTO expression induced by cagA
+
H. pylori infection was abolished by JUN knockdown (Figure 3N, Supplementary Figure S3H). Consistent with these findings, JUN knockdown also eliminated the metastatic and invasive capabilities of GC cells induced by cagA
+
H. pylori infection (Supplementary Figure S3I‐J). Collectively, our findings confirm that cagA
+
H. pylori infection promotes FTO transcription by activating JUN in GC cells.

3.4
CagA‐induced FTO elevation further promoted EMT in GC cells
To identify candidate targets of the CagA‐FTO axis, MeRIP‐seq and RNA‐seq were conducted on MKN45 cells with or without cagA
+
H. pylori infection (Supplementary Figure S4A). MeRIP‐Seq results indicated that mRNAs from the cagA
+
H. pylori‐infected group exhibited a lower number of m6A peaks compared to the non‐infected control group. (Supplementary Figure S4B). Notably, a total of 1,622 genes with m6A altered peaks exhibited downregulation, while 2,151 genes showed upregulation in expression after cagA
+
H. pylori infection (Supplementary Figure S4C). GO analysis indicated that genes with altered m6A peaks and differential expression after cagA
+
H. pylori infection were substantially enriched in processes related to cell polarity, including cytoskeleton and microtubule organization (Supplementary Figure S4D). By intersecting data from our RNA‐seq and MeRIP‐seq with RNA‐seq data from the GSE178697, we identified 336 genes that displayed decreased m6A modification and altered expression following FTO knockdown and H. pylori infection (Supplementary Figure S4E), suggesting their potential as downstream targets of the CagA‐FTO axis. These candidate targets were notably enriched in the EMT pathway, as assessed via GSVA (Supplementary Figure S4F). The microsatellite stable/epithelial‐to‐mesenchymal transition (MSS/EMT) subtype of GC, known for its aggressive nature and unfavorable prognosis [34], showed significantly elevated FTO expression levels compared to other subtypes in our analysis of the GSE62254 dataset (Supplementary Figure S4G). The EMT score, based on the expression differences of known epithelial and mesenchymal markers [35], showed a positive correlation with FTO expression in GC tissues from the TCGA‐STAD cohort (Supplementary Figure S4H). GSEA further confirmed that the EMT molecular signature was notably enriched in GC samples with high FTO expression compared to those with low FTO expression (Supplementary Figure S4I).
The expression of mesenchymal markers, including N‐cadherin, Snail, and Vimentin, was substantially reduced in the FTO knockdown group compared to the control group, while E‐cadherin expression was increased (Supplementary Figure S4J). Conversely, overexpression of FTO, but not its mutant forms, significantly elevated the expression of mesenchymal markers (Supplementary Figure S4K). Moreover, FTO knockdown in GC cells reversed the EMT process induced by cagA
+
H. pylori infection (Supplementary Figure S4L). Overall, these findings suggest that elevated FTO expression following cagA
+
H. pylori infection may activate EMT, which may account for the pro‐oncogenic effects of the bacterial infection.

3.5
HBEGF served as the primary downstream target of the CagA‐FTO axis
To deepen our comprehension of the CagA‐FTO axis's regulatory role in EMT, we intersected the 336 putative downstream targets with the EMT‐associated genes from the EMTome database (Figure 4A). HBEGF, FKBP5, EPB41L3, EPHA8, PGF, KRT18, MAGED1, PHLDA2, WNT3A, and TIAM1 were determined to be regulated by FTO. Further analysis identified HBEGF as the gene exhibiting the most striking reduction in response to FTO knockdown (Supplementary Figure S5A). In addition, HBEGF expression was notably elevated in GC cells infected with cagA
+
H. pylori compared to those infected with ΔcagA H. pylori or uninfected (Figure 4B). Our Western blot analysis further confirmed that FTO knockdown reduced HBEGF expression in GC cells (Figure 4C). Conversely, only FTO‐WT‐oe, rather than mutant forms FTO‐mut1‐oe and FTO‐mut2‐oe, substantially increased HBEGF expression (Figure 4D). Furthermore, the elevation of HBEGF induced by cagA
+
H. pylori was reversed by FTO knockdown (Figure 4E), establishing HBEGF as a primary downstream target of the CagA‐FTO axis. A positive correlation was observed between HBEGF levels and the EMT score in the vast majority of tumor types (Supplementary Figure S5B). GC patients with elevated HBEGF expression also showed poor OS (Supplementary Figure S5C). ​Studies have implicated secreted HBEGF in promoting EMT [36, 37, 38]. We confirmed that cagA
+
H. pylori infection and FTO overexpression significantly increased HBEGF secretion, while FTO knockdown reduced HBEGF secretion (Figure 4F‐H).

HBEGF knockdown markedly reduced GC cell metastasis and invasion (Supplementary Figure S5D‐E) and negated the pro‐EMT and pro‐metastatic effects induced by cagA
+
H. pylori infection (Figure 4I, Supplementary Figure S5F‐G). To further uncover the m6A modification sites on HBEGF mRNA regulated by FTO, we utilized the SRAMP database to predict four m6A modification sites on HBEGF mRNA with “very high confidence” (Supplementary Figure S6A). On the basis of these four sites, we designed specific primers and further confirmed them via MeRIP‐qPCR. Results showed that FTO knockdown significantly elevated the level of m6A modification at sites 418, 448, and 1304 in MKN45 cells (Figure 4J), and at sites 418 and 448 in AGS cells (Supplementary Figure S6B). Consequently, we infer that sites 418 and 448 may be the m6A sites specifically targeted by FTO on HBEGF mRNA. Moreover, the m6A modifications at these sites were markedly reduced by the overexpression of FTO‐WT, compared with NC and FTO‐mut1/FTO‐mut2 (Figure 4K, Supplementary Figure S6C). Taken together, these findings indicated that FTO can specifically target m6A sites on HBEGF mRNA, modulating its m6A modification level.

3.6
FTO increased the stability of HBEGF mRNA via YTHDF2
m6A readers have been reported to control the stability, translation, and shearing of RNA [25, 39]. To identify m6A readers that target HBEGF mRNA, we performed mass spectrometry analysis following an RNA pull‐down assay (Supplementary Tables S7‐S8). Among three replicates, 135 candidate proteins were screened out. After exhaustive research in specialized databases and literature, only YTHDF2 was reported as an m6A “reader” among these proteins [25]. Thus, we suggest that YTHDF2 may be a candidate for binding to HBEGF mRNA (Figure 5A). RIP and RNA pull‐down assays confirmed that YTHDF2 binds to HBEGF mRNA (Figure 5B‐C). YTHDF2, a protein known to target m6A sites and decrease mRNA stability [25], exhibited a significant elevation in HBEGF expression following YTHDF2 knockdown, as compared to the control group (Figure 5D, Supplementary Figure S7A). Furthermore, YTHDF2 knockdown significantly enhanced HBEGF mRNA stability (Figure 5E), whereas FTO knockdown reduced it (Figure 5F). In addition, only the overexpression in cells with FTO‐WT‐oe, but not with FTO‐mut1‐oe and FTO‐mut2‐oe, was found to increase HBEGF mRNA stability (Figure 5G, Supplementary Figure S7B).
​An inverse relationship was identified between the expression levels of YTHDF2 and those of HBEGF and FTO (Figure 5H, Supplementary Figure S7C‐E). Furthermore, elevated YTHDF2 expression in GC tissue was associated with better patient outcomes (Figure 5I). YTHDF2 knockdown in GC cells with stable FTO knockdown rescued HBEGF expression (Figure 5J, Supplementary Figure S7F). To ascertain whether FTO and YTHDF2 modulate HBEGF expression through m6A modification, we stably expressed wild‐type HBEGF with a Flag tag and HBEGF mutations at the 418 or 448 m6A sites (Figure 5K, Supplementary Figure S7G), as previously validated (Figure 4J‐K). Neither FTO nor YTHDF2 influenced HBEGF expression after mutation of the m6A sites in HBEGF (Figure 5L‐N, Supplementary Figure S7H‐I). These results indicate that YTHDF2 exploits the 418 and 448 m6A sites to regulate HBEGF mRNA stability.

3.7
MA treatment impeded the progression of GC following H. pylori eradication
The “hit‐and‐run” hypothesis posits that H. pylori eradication may not prevent the progression of GC [18, 19]. In line with this hypothesis, we observed that although CagA was substantially eliminated in GC cells following antibiotic treatment, the levels of FTO and HBEGF did not revert to baseline but remained elevated compared to the uninfected group (Figure 6A, Supplementary Figure S8A). To explore this phenomenon, we constructed a DOX‐CagA‐tet‐on system, which allows DOX‐inducible overexpression of cagA, as previously described [18]. Mimicking chronic cagA
+
H. pylori infection with intermittent delivery of CagA via T4SS, we induced CagA by 20 cycles of DOX (Figure 6B). This repeated induction of CagA promoted FTO and HBEGF expression, and even after 48 hours of cessation of CagA induction, their expression levels remained elevated (Figure 6C‐D, Supplementary Figure S8B), suggesting that repeated exposure with cagA
+
H. pylori could induce irreversible epigenetic alterations.
MA, a nonsteroidal anti‐inflammatory drug (NSAID) approved by the U.S. Food and Drug Administration (FDA) [40], has been reported to selectively inhibit the m6A demethylation activity of FTO without affecting AlkB homolog 5 (ALKBH5) [41]. Treatment with MA increased the m6A level (Figure 6E) and significantly suppressed HBEGF expression (Figure 6F‐G), as well as the levels of N‐cadherin, Vimentin and Snail, while upregulating E‐Ca (Figure 6H). Consistently, MA disrupted the migration and invasion capabilities of GC cells (Figure 6I, Supplementary Figure S8C). These effects were also verified in human GC organoids, where MA treatment inhibited the EMT process (Supplementary Figure S8D‐G). To assess the combined therapeutic effect of MA and antibiotics on GC, organoids were treated with MA and kanamycin following infection with cagA
+
H. pylori. Eradication of H. pylori post‐infection mitigated the EMT process, characterized by increased E‐cadherin and decreased N‐cadherin, Snail, and Vimentin; importantly, subsequent MA treatment further inhibited EMT activation (Figure 6J‐N). Collectively, these findings indicate that cagA
+
H. pylori infection may induce irreversible alterations in the oncogenic pathway through m6A modification, and MA therapy potentially blocks CagA‐induced EMT and the metastatic capacity of GC cells.

3.8
CagA targeted FTO to promote GC metastasis in vivo
We extended our investigation to in vivo models to assess the impact of FTO on GC metastasis. A mouse model of tumor metastasis was established by tail vein inoculation of GC cells. Notably, FTO knockdown significantly inhibited lung metastasis of GC cells (Figure 7A‐B, Supplementary Figure S8H). Conversely, overexpression of FTO‐WT, but not FTO‐mut1 and FTO‐mut2, led to an increase in the number of lung metastases (Figure 7C‐D). To explore the therapeutic potential of MA in reversing CagA‐promoted GC metastasis in vivo, CagA‐tet‐on cells were inoculated via the tail vein, followed by treatment with DOX and MA for 30 days (Figure 7E). The results showed that DOX‐induced CagA significantly promoted the lung metastasis of GC cells. However, MA treatment effectively abolished the contribution of CagA on cell metastasis in the lungs (Figure 7F‐G).
To further validate the above results in GC patients, expression of cagA and FTO were analyzed in two TMAs of GC samples. FTO expression was found to be noticeably elevated in GC tissues (Figure 7H). Importantly, FTO levels were considerably higher in CagA‐positive GC tissues compared with CagA‐negative GC tissues (Figure 7I). Moreover, the presence of CagA and increased FTO expression in GC patients were correlated with poor outcomes (Figure 7J). Collectively, the findings from both mouse models and clinical GC patients underscore the role of the CagA‐FTO axis in driving GC progression.

4
DISCUSSION
Although the interaction between H. pylori and the host has a significant impact on the development of GC, the underlying mechanisms remain largely unclear [42, 43]. Our research provided evidence that CagA delivery by H. pylori reduced total m6A levels and promoted FTO expression in GC cells, which facilitated the metastatic potential of GC. We clearly demonstrated that CagA reduced m6A modification in biomolecules linked to EMT, particularly HBEGF, by increasing FTO transcription through JUN. HBEGF expression was elevated as a consequence of the reduced m6A modification, which prevented YTHDF2‐dependent mRNA from degrading. Importantly, the results we gathered highlight MA's potential effectiveness as an FTO antagonist to prevent the progression of GC. MA administration was successful in preventing EMT in GC cells, and it showed efficacy in inhibiting GC metastasis in both animal models and human GC organoids. The results not only indicate that FTO inhibitors are potential drugs for GC treatment but also, more importantly, provide novel insights into the “hit‐and‐run” mechanism of CagA‐induced GC development from an epigenetic perspective (Figure 8).
As the most abundant form of RNA modification, m6A is dynamically altered in response to environmental and physiological influences and plays a crucial role in the development of multiple types of cancers [23, 44]. It is now established that several microbes have the capacity to modulate the m6A landscape of their hosts [20], with these epigenetic alterations supporting microbial replication and influencing the host's disease process [45, 46]. Abnormal expression of FTO has been observed across multiple tumor types [47, 48, 49]. Additionally, FTO has been implicated in the malignant development of tumors caused by environmental alterations [21, 50]. Our results also demonstrated that CagA, a bacterial protein delivered by H. pylori, upregulated FTO expression and facilitated the EMT process. Both the results of previous studies and our research suggest that FTO is a responsive factor to environmental stress and plays a significant role in tumor progression.
The process of tumor metastasis, which involves a series of steps, initially stages with EMT [51, 52]. The activation of EMT is strongly interrelated with tumor metastasis and invasion [51, 53]. It was revealed in our study that cagA
+
H. pylori triggers the EMT process in an FTO‐dependent manner. Several EMT‐related genes (HBEGF, FKBP5, EPB41L3, EPHA8, PGF, KRT18, MAGED1, PHLDA2, WNT3A, and TIAM1) were screened and determined to be regulated by FTO. Among these genes, HBEGF is capable of inducing EMT in a variety of epithelial cells [36, 54, 55]. Previous studies had shown that the extracellular domain of HBEGF was cleaved after synthesis, resulting in the shedding of biologically active soluble HBEGF [56]. An increase in secreted HBEGF was shown to induce EMT and promote epithelial cell invasion, whereas membrane‐bound HBEGF failed to induce EMT [36, 37, 38]. Our findings confirmed that HBEGF expression and secretion, determining factors in EMT promotion, were regulated by the CagA‐FTO axis.
Clinical evidence is supportive of the continued necessity for H. pylori eradication in post‐surgical GC patients [12, 13, 15]. However, it has been indicated by several studies that GC progression may persist despite the eradication of H. pylori [16, 17]. This finding may stem from the fact that genomic and epigenetic alterations induced by cagA
+
H. pylori infection are not entirely reversed by eradication. The intricate mechanism behind this “hit‐and‐run” effect remains unclear. Interestingly, it was revealed in our results that CagA was largely eliminated after antibiotic treatment, while the CagA‐induced expression of FTO and HBEGF remained significantly elevated compared to pre‐infection levels. This observation was validated through a CagA‐tet‐on system designed to mimic chronic cagA
+
H. pylori infection. Insights into a possible molecular mechanism for the CagA‐induced “hit‐and‐run” phenomenon were offered by our results, highlighting m6A modification as a key epigenetic factor. The importance of mitigating the epigenetic changes induced by CagA was underscored by these findings. Consequently, restoring m6A homeostasis may be a potential therapeutic method for treating GC with cagA
+
H. pylori infection. Satisfactory outcomes have been yielded by preclinical experiments on the discovery of FTO inhibitors [47, 57, 58]. MA, as an FTO‐specific inhibitor that does not affect the function of ALKBH5 [41], was previously approved by the FDA for the treatment of osteoarthritis [40]. Guided by these findings, MA was selected to counteract the effects of CagA‐induced FTO upregulation. Our data fully confirmed that a combination of MA and antibiotics could significantly impede the progression of GC.
While our study provides evidence that infection with cagA
+
H. pylori may lead to irreversible epigenetic alterations, several limitations warrant discussion. Our findings, initially observed in cell line models, were corroborated using a CagA‐tet‐on system. However, the generalizability of these results to clinical settings needs further validation through patient‐involved interventional studies. Moreover, the mechanisms underlying the epigenetic reprogramming induced by CagA, which appear to be irreversible, require a more comprehensive investigation. Animal studies typically employ models of lung metastasis to assess the metastatic potential of GC [59, 60], employing a broader range of in vivo models could potentially reinforce the conclusions drawn from our experiments. Comorbidities, including preexisting conditions and concurrent infections, may synergistically accelerate the progression of GC [5, 61, 62]. Comprehensively analysis of the role of H. pylori infection and these comorbidities in promoting GC progression is required for future studies. Moreover, no studies have linked the use of MA to GC development, although epidemiologic data suggest that NSAID use reduces the incidence of noncardiac GCs [63, 64]. The identification of further studies on the clinical use of MA in GC patients will be interesting because our study confirmed the efficacy of MA in preventing GC progression.

5
CONCLUSIONS
This study elucidates the molecular mechanism of FTO's role in CagA‐promoted GC progression, and provides new evidence that CagA‐induced GC progression through a “hit‐and‐run” mechanism, characterized by irreversible epigenetic changes even after the bacterium's clearance. Our results provide strong evidence for the further clinical exploration of strategies aimed at blocking FTO activity. Such interventions hold promise for halting the CagA‐induced “hit‐and‐run” effects and potentially preventing the progression of GC.

AUTHOR CONTRIBUTIONS
Shiming Yang and Yufeng Xiao designed the project. Bing He, Yiyang Hu, Yuyun Wu, Chao Wang, Limin Gao, Chunli Gong, Zhibin Li, Nannan Gao and Huan Yang performed the experiments. Yufeng Xiao and Bing He analyzed the data and prepared the figures and tables. Bing He, Limin Gao and Chao Wang collected all of the tissue samples. Bing He, Yufeng Xiao and Shiming Yang wrote the manuscript and all of the authors proofed it. Shiming Yang and Yufeng Xiao conceived the project and supervised and coordinated all aspects of the work.

CONFLICT OF INTEREST STATEMENT
The authors declare that they have no competing interests.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE
This study was approved by the Medical Ethics Committee of Xinqiao Hospital of Army Medical University (Approved No. 2024‐083‐01), and by Shanghai Outdo Biotech Co. Ltd. (Approved No. YBM‐05‐02). All animal studies were conducted in compliance with protocols approved by the Laboratory Animal Welfare and Ethics Committee of Army Military Medical University (Approved No. AMUWEC20235107).

Supporting information
Supporting Information