DNA methylation in gastric precancerous lesions: Molecular mechanisms and clinical translation

Abstract

DNA methylation constitutes a central epigenetic mechanism driving the initiation and progression of gastric precancerous lesions (GPL). This review synthesizes current evidence on the molecular basis of methylation dysregulation in GPL, establishing a mechanistic framework encompassing upstream drivers (Helicobacter pylori infection, Epstein-Barr virus, host aging, metabolic reprogramming) and downstream biological consequences (tumor suppressor gene silencing, Wnt/β-catenin activation, epithelial-mesenchymal transition, and microenvironment remodeling). Clinically, DNA methylation biomarkers detected in tissue, blood, stool, gastric juice, and emerging breath samples demonstrate substantial utility for early screening, risk stratification, prognostic prediction, and therapeutic response monitoring. Furthermore, targeted strategies modulating methylation-regulating enzymes and methylation-guided chemoprevention represent promising precision intervention paradigms. Despite challenges including epigenetic heterogeneity and translational barriers, integrating multi-omics technologies and developing multimodal liquid biopsy assays will propel the clinical implementation of methylation-based management for GPL.

Key Words: DNA methylation; Gastric precancerous lesions; Epigenetic biomarkers; Helicobacter pylori; Intestinal metaplasia; Risk stratification; Early detection of cancer; Epigenetic therapy; Chemoprevention; Molecular mechanisms

Core Tip: DNA methylation is a pivotal epigenetic driver in gastric precarcinogenesis, orchestrating the stepwise progression from precancerous lesions to cancer. Clinically, methylation biomarkers enable non-invasive early detection, risk stratification, and prognostic assessment. Therapeutically, targeting methylation-regulating enzymes and implementing methylation-guided chemoprevention represent promising precision intervention strategies.

INTRODUCTION

Gastric cancer (GC) remains one of the most prevalent and lethal malignancies worldwide, according to GLOBOCAN estimates of cancer incidence and mortality[1]. Often diagnosed at an advanced stage, GC carries a poor prognosis and continues to impose a considerable global health burden. The pathogenesis of GC typically follows a stepwise progression from normal gastric mucosa to precancerous lesions—including chronic atrophic gastritis, intestinal metaplasia (IM), and dysplasia—and eventually to adenocarcinoma. These precancerous stages represent a critical phase in the natural history of GC[2]. Pan-cancer analyses have demonstrated that gene promoter hypermethylation is an early hallmark of carcinogenesis, with tumor-type-specific methylation patterns forming the molecular basis for DNA methylation-based early detection strategies[3]. While genetic mutations and chromosomal alterations remain central drivers of carcinogenesis, accumulating evidence highlights the critical role of epigenetic modifications, particularly DNA methylation, in initiating and advancing gastric precancerous lesions (GPL). Comprehensive reviews of gastric epigenetics underscore that DNA methylation, along with histone modifications and non-coding RNAs, contributes to both initiation and progression of gastric carcinogenesis. These insights have accelerated the development of epigenetic biomarkers and targeted therapeutic approaches that are rapidly moving toward clinical translation[4]. Mechanistically, DNA methylation involves the addition of a methyl group to cytosine residues within CpG islands, particularly in promoter regions. This modification frequently results in transcriptional silencing of tumor suppressor genes (TSGs), thereby promoting oncogenic transformation. Environmental factors, most notably Helicobacter pylori (H. pylori) infection, induce widespread methylation alterations. These changes contribute to the molecular pathogenesis of precancerous lesions while simultaneously offering diagnostic and prognostic opportunities for early detection, patient risk stratification, and disease monitoring. Furthermore, the reversibility of epigenetic modifications presents novel therapeutic opportunities. Accordingly, this review synthesizes current evidence on the role of DNA methylation in driving GPL, explores its clinical applications in screening and risk prediction, and discusses emerging therapeutic strategies targeting the epigenetic machinery, while also highlighting ongoing challenges and future research directions beyond DNA methylation.

DNA METHYLATION-DRIVEN MOLECULAR MECHANISMS IN GPL

Current evidence indicates that aberrant DNA methylation is not an isolated event, but rather a core mechanism that, driven by multiple factors and through disruption of key cellular pathways, progressively promotes lesion evolution.

Upstream drivers: Exogenous and endogenous sources of methylation imbalance

Aberrations in DNA methylation are triggered by a combination of exogenous stimuli and endogenous factors, which constitute the initiating and promoting conditions for lesion development.

Exogenous drivers: H. pylori infection is a well-established environmental factor that initiates gastric carcinogenesis by inducing widespread epigenetic alterations. Genome-wide bisulfite sequencing revealed that SSTR2 promoter hypermethylation occurs in H. pylori-infected aged gastric mucosa, driving the establishment of a pro-inflammatory microenvironment through gene silencing. Functional studies confirmed that loss of SSTR2 promotes proliferation, whereas its overexpression suppresses tumorigenesis, demonstrating causal epigenetic regulation in the initiation of GC[5]. Purkait et al[6] further showed that DNMT-1/3B overexpression in H. pylori-associated gastritis correlates with IM and that crosstalk between DNMTs and the histone modifier EZH2 synergistically promotes gastric carcinogenesis, underscoring the importance of epigenetic interplay in precancerous progression. Mechanistic studies demonstrate that H. pylori infection induces DNMT1-mediated MTTP promoter methylation, which suppresses GPX4 expression and activates ferroptosis, thereby promoting progression from atrophic gastritis to GC[7]. Genome-wide analyses further revealed that concurrent CpG island hypermethylation and non-CpG hypomethylation in gastric mucosa after H. pylori eradication promote oncogene activation, maintaining residual cancer risk[8]. Additionally, H. pylori infection activates the JAK/STAT pathway via IL6/IL10 signaling and induces SOCS1 promoter hypermethylation, leading to its silencing and thereby driving inflammation-associated gastric carcinogenesis[9]. Consistent findings indicate that H. pylori-induced dysregulation of SOCS1/3 expression through DNA methylation suppresses immune responses, sustains chronic inflammation, and promotes gastric carcinogenesis[10]. Clinical studies have also demonstrated a significant association between H. pylori infection and CDH1 promoter hypermethylation in gastric mucosa, observed in both gastritis and GC tissues, suggesting that this alteration is an early event in gastric carcinogenesis[11]. Evidence suggests that H. pylori infection may induce the earliest precancerous alterations, including aberrant DNA methylation, in the gastric mucosa during childhood, particularly in children carrying highly virulent strains[12]. In addition, Epstein-Barr virus (EBV) infection serves as another significant exogenous driver. In EBV-associated GC, LMP1 recruits DNMT1 to hypermethylate RASSF10, silencing its expression and promoting proliferation, a pathway that may be therapeutically targetable[13].

Endogenous drivers: Host aging provides a susceptible background for the development of GPL. Theoretical models propose that age-related accumulation of stochastic DNA methylation increases epigenetic entropy, driving a transition of normal cells toward a cancer stem cell-like state, which may provide a novel framework for predicting the risk of GPL[14]. Organoid studies further revealed that methylation of the Dkk3 gene in aged gastric epithelial cells leads to reduced expression, enhanced Wnt signaling, and Tbx3-mediated suppression of senescence, thereby promoting accelerated proliferation. This alteration has also been confirmed in human GPL and carcinomas[15]. Epigenetic-metabolic crosstalk has emerged as a critical component of gastric precancerous progression. Promoter hypermethylation of genes such as CDX2 drives IM by silencing tumor suppressors, while interacting with methionine metabolic reprogramming. These combined alterations highlight potential therapeutic opportunities for precision-based interventions in GPL[16].

Downstream core effects: Biological consequences driven by methylation alterations

The specific methylation changes initiated by upstream drivers converge into several distinct oncogenic biological effects through the silencing of key genes.

Dysregulation of key signaling pathways and inactivation of TSGs: Promoter hypermethylation that leads to the inactivation of numerous TSGs represents the most direct molecular mechanism driving lesion progression. These inactivated genes have diverse functions and collectively dismantle cellular homeostasis. DNMT3a-driven FBP2 promoter hypermethylation silences glycolysis suppression, facilitating gastric carcinogenesis and offering potential as a prognostic biomarker[17]. Methylation-induced silencing of Dkk3 associated with aging also leads to enhanced Wnt signaling[15]. Integrated analyses demonstrate that hypermethylation-induced transcriptional silencing of the CDH1 promoter represents a universal epigenetic event in GC, correlating with epithelial-mesenchymal transition (EMT) and stemness pathway activation[18]. A study in a southern Chinese cohort revealed promoter hypermethylation of EYA4 in gastric cardia IM, resulting in reduced protein expression and suggesting a role for epigenetic silencing in premalignant progression at the gastric cardia[19]. Hypermethylation-induced silencing of ZNF677 drives its transcriptional silencing during gastric precancerous progression and carcinogenesis, with this tumor suppressor inactivation promoting tumorigenesis through regulating cellular processes like proliferation[20]. Validated by mRNA/protein assays, a study revealed that promoter hypermethylation of DHRS3 in GC significantly downregulates its expression. Ectopic DHRS3 expression suppresses tumor proliferation and migration, and hypermethylation at CpG9.10 correlates with shortened patient survival, implicating epigenetic inactivation of this tumor suppressor[21].

Fundamental alterations in cellular identity and fate: DNA methylation reprogramming alters the fate of gastric epithelial cells and underlies the morphological transformation during the precancerous stage. On one hand, DNA methylation reprogramming drives IM. Research has confirmed that IM cells exhibit widespread promoter hypermethylation (an epigenetic footprint), and NOS2 overexpression accelerates epigenetic instability by upregulating DNMT activity, establishing the precancerous potential of IM[22]. GPL, particularly IM, are characterized by epigenetic reprogramming of developmental genes. Genome-wide methylation profiling identified more than 38000 differentially methylated probes in gastric cardiac IM, with promoter CpG island hypermethylation frequently targeting developmental genes. For example, HOXA5 hypermethylation correlates with transcriptional silencing and may serve as a biomarker of gastric precancerous pathogenesis[23]. Spatial multi-omics analysis reveals that incomplete IM displays unique intergenic hypermethylation and lineage plasticity, resembling GC at the molecular level and can be regarded as a genuine precancerous lesion[24]. On the other hand, DNA methylation reprogramming promotes EMT and the acquisition of invasive potential. As described, methylation-induced silencing of genes such as CDH1 directly drives EMT[18]. KDM6A downregulation activates Wnt/β-catenin signaling, enhancing EMT and correlating with poor prognosis[25].

Epigenetic remodeling of the tumor microenvironment: Methylation alterations not only affect epithelial cells but also participate in shaping a pro-carcinogenic microenvironment. H. pylori infection drives inflammation-associated carcinogenesis by inducing hypermethylation and silencing of genes like SOCS1, thereby suppressing the negative feedback of the JAK/STAT pathway and sustaining persistent inflammatory signaling[9,10]. Patients with nodular gastritis exhibit pronounced hypermethylation of TSGs such as CDH1 and DAPK1, coupled with reduced expression of DNA demethylation genes such as TET2 and IDH1, which suggests that epigenetic imbalance drives malignant transformation[26]. Aberrant overexpression of FAM64A, negatively regulated by methylation, emerges as early as the IM stage and promotes proliferation and invasion via the EGFR/Akt pathway[27]. Stepwise increases in methylation aberrations have been observed across gastric cardia precancerous lesions, progressing from type I to type III IM and peaking in intraepithelial neoplasia (IEN), with shared promoter hypermethylation between type III IM and IEN[28].

A dynamic integrative model: The cascade from initiation to malignant progression

In summary, the role of DNA methylation in driving GPL is not a linear process but rather a multi-factorial, multi-step dynamic cascade involving feedback loops (Figure 1). During the initiation phase, exogenous triggers, against a susceptible endogenous background, induce initial, specific methylation silencing of TSGs by directly regulating methyltransferases or provoking chronic inflammation. In the propagation and amplification phase, these initial methylation changes disrupt key pathways, leading to fundamental alterations in cell fate. These abnormal cells can further modify the local microenvironment through secreted factors. Concurrently, endogenous changes such as metabolic reprogramming provide substrate support for sustained methylation modifications. The aberrant methylation state itself may form a positive feedback loop by upregulating DNMT activity or interacting with histone modifications, leading to increased epigenetic instability and heterogeneity, thereby accelerating clonal evolution[22]. Regarding outcome, the interaction between different combinations of drivers and the host genetic background generates characteristic methylation profiles (e.g., global hypermethylation in EBV-positive GC, MLH1 promoter hypermethylation in microsatellite instability subtypes), ultimately determining the clinicopathological heterogeneity of the lesions[29,30]. A large-scale international study discovered that IM harbors specific driver mutations such as in ARID1A, a characteristic SBS17 mutational signature, and that clonal hematopoiesis may promote the progression from IM to GC by modulating mucosal immunity[31].

Figure 1 Molecular mechanisms of DNA methylation driving the progression of gastric precancerous lesions. H. pylori: Helicobacter pylori; EMT: Epithelial-mesenchymal transition.

CLINICAL APPLICATIONS OF DNA METHYLATION MARKERS IN THE MANAGEMENT OF GASTRIC PRECANCEROUS LESIONS

DNA methylation analysis in clinical screening and diagnosis

DNA methylation markers, particularly those detectable in non- or minimally invasive samples, have shown considerable potential for early screening and differential diagnosis of GC and its precancerous lesions, often outperforming conventional diagnostic methods.

Stool-based testing is a commonly used non-invasive approach. A study demonstrated that aberrant hypermethylation of SDC2 and TERT in fecal DNA represents a biomarker for GC screening. While the single marker SDC2 achieved an accuracy of 71.2%, a combined panel incorporating SDC2/TERT methylation and fecal hemoglobin substantially improved sensitivity for stage I GC (78.6%), underscoring the role of promoter hypermethylation as an epigenetic driver of early gastric carcinogenesis[32].

Blood-based assays also provide significant complementary value. A prospective cohort study showed that combining blood-based methylation markers mSEPT9 and mRNF180 with the CA724 protein significantly enhanced GC detection sensitivity. Importantly, hypermethylation of the DOK7 gene CpG island in peripheral blood leukocytes was detected in 66.0% of patients with IM and 88.1% of those with GC, highlighting its potential as a noninvasive biomarker for GPL[33]. Genome-wide cfDNA profiling identified a 133-marker panel for gastric precancerous lesion risk assessment and a 49-marker panel integrated with mutation data for improved GC detection, emphasizing the role of methylation signatures in risk stratification and early diagnosis[34]. Circulating methylated RPRM cfDNA levels increased progressively with gastric precancerous development, enabling monitoring of therapeutic responses, and when combined with the pepsinogen ratio, enhanced early lesion detection[35].

Tissue samples are considered the diagnostic “gold standard”, as their methylation analysis can provide the most direct information regarding the nature of the lesion. An integrated methylome-transcriptome study developed a methylation marker panel specific to gastric cardia adenocarcinoma that achieved an area under the curve of 0.917 for distinguishing high-grade dysplasia/carcinoma from low-grade or normal tissues[36]. Targeted sequencing using the OPERA_MET-A panel has enabled deep CpG methylation analysis across multiple genes, offering a sensitive approach for detecting early-stage gastric lesions[37].

Beyond the mainstream sample types discussed above, gastric juice or gastric lavage fluid serves as a “local liquid biopsy” in closer proximity to the lesion site, potentially offering higher lesion specificity. Research indicates that gastric juice contains secretions from the gastric mucosa and may thus reflect changes associated with GC development stages. Potential biomarkers for GC screening identified in gastric juice may mirror alterations linked to the developmental stages of GC[38]. Additionally, completely non-invasive exhaled breath samples represent another important exploratory avenue, although research on related epigenetic markers remains in its early stages. Hong et al[39] found that a simple model based on the urea breath test, pepsinogen, and gastrin-17 demonstrated high predictive value for GPL, holding promise as a novel method for clinically identifying high-risk populations.

DNA methylation detection methods from different sample sources each have their own emphases in terms of performance and application (Table 1). Future research could focus on developing detection methods utilizing novel samples such as gastric juice and employing multi-modal combination strategies to overcome the limitations of single methods, ultimately achieving early diagnosis with higher sensitivity and specificity.

Table 1 Comparison of DNA methylation detection methods for screening and diagnosis of gastric precancerous lesions.

Sample type	Performance characteristics	Core advantages	Main limitations
Tissue	High	Directly reflects molecular features at the lesion site; allows simultaneous pathological diagnosis	Invasive sampling; not suitable for large-scale, repeated screening
Blood	Moderate-high	Completely non-invasive; facilitates repeated sampling and dynamic monitoring	Sensitivity may be limited for very early or focal lesions; relatively high cost
Stool	Moderate	Completely non-invasive; convenient sample collection and high patient compliance	Results may be influenced by factors such as gut microbiota
Gastric juice	To be systematically validated (theoretically potentially high)	A local liquid biopsy closer to the lesion and may enrich lesion-specific signals	Sampling requires intubation via the mouth, reducing comfort
Breath test	Early research stage	Completely non-invasive with extremely high patient acceptance	Technological platforms are in early development; research on related epigenetic markers is scarce

DNA methylation-based risk stratification and prognostic assessment models

Risk stratification models based on DNA methylation biomarkers enhance the predictive capability for the progression of GPL, patient prognosis, and the risk of recurrence/metastasis, thereby providing critical decision-making support for the full-cycle personalized management of GC.

Dynamic methylation patterns are closely linked to lesion progression and future cancer development. A longitudinal study of gastric cardia precancerous lesions found that DNA methylation abnormalities progressively accumulate from type I IM to type III IM and further to IEN, with type III IM and IEN sharing specific methylation features[28]. Longitudinal studies have shown that IM recapitulates GC methylation profiles, with elevated RPRM and ZNF793 methylation independent of H. pylori infection, suggesting their role as biomarkers of progression risk[40]. Persistent hypermethylation of miR-124a-3 and EMX1 following H. pylori eradication has been associated with precancerous mucosal changes and increased metachronous GC risk[41]. Methylomic profiling also revealed progressive promoter hypermethylation of C11orf87 during precancerous progression, which correlated with favorable outcomes, indicating its utility as a prognostic biomarker[42]. A prospective cohort demonstrated that baseline MOS methylation ≥ 34.82% in gastric antrum significantly predicted metachronous recurrence after endoscopic resection, outperforming OLGA/OLGIM staging in predictive performance[5]. Statistical models incorporating DNA methylation data showed that the cumulative number of stem cell divisions was a strong predictor of IM risk, and integration of gastric atrophy status further improved stratification accuracy[43]. Furthermore, risk stratification extends to prognosis and metastatic potential.

DNA methylation biomarkers demonstrate significant value in predicting the survival outcomes of GC patients. Studies indicate that circulating methylated THBS1 DNA can serve as a novel marker for predicting peritoneal dissemination in GC, offering a non-invasive tool for assessing metastasis risk and poor prognosis[44]. Furthermore, the methylation status of genes such as NCOR2, PARK2, and ZSCAN12 in tumor tissue is significantly correlated with the density of tumor-infiltrating lymphocytes, suggesting their potential impact on prognosis by influencing the tumor immune microenvironment[45]. These findings collectively indicate that methylation signatures, whether derived from tissue or liquid biopsies, can provide crucial information for evaluating patients' overall survival and disease-free survival.

Methylation markers are crucial for predicting therapeutic response and guiding personalized treatment. In the context of chemotherapy, promoter methylation of the SLFN11 gene leads to its transcriptional silencing and is associated with resistance to DNA-damaging agents in GC, whereas demethylating interventions may restore chemosensitivity[46]. This provides a basis for utilizing methylation status to predict chemotherapy efficacy. Regarding immunotherapy, functional CpG site methylation in the PD-L1 gene promoter has been identified as a novel epigenetic biomarker for primary GC[47].

Beyond the aforementioned prediction of peritoneal dissemination[44], alterations in DNA methylation also play a role in predicting other metastatic routes and assessing overall metastatic potential. Multi-omics analysis has revealed cuproptosis-related methylation features in GC and constructed a risk model that can influence the prognosis of stomach adenocarcinoma[48]. Furthermore, studies have confirmed that DNA methylation variability can serve as a field cancerization marker for identifying adenoma patients at risk of developing metachronous advanced colorectal lesions[49]. This concept is equally applicable to the field of GC, suggesting that localized or systemic abnormal methylation patterns may serve as macroscopic indicators for predicting the risk of multifocal gastric lesions or distant metastasis.

Figure 2 illustrates the clinical translation and application of DNA methylation biomarkers in GPL.

Figure 2 Clinical translation and application of DNA methylation biomarkers in gastric precancerous lesions. cfDNA: Cell-free DNA; AI: Artificial intelligence.

THERAPEUTIC INTERVENTIONS AND PREVENTION STRATEGIES BASED ON DNA METHYLATION MARKERS

Therapeutic strategies targeting methylation-regulating enzymes

Targeting DNA methylation-regulating enzymes represents a significant research direction in the epigenetic therapy of GC, with core strategies focusing on reversing pathogenic gene silencing or exploiting tumor-specific methylation defects. Currently, several preclinical studies provide proof-of-concept in developing novel modulators and overcoming therapeutic resistance.

New DNMT3B inhibitors discovered through computer-aided screening have been shown to inhibit aberrant methylation and tumorigenesis in preclinical studies[50]. Research targeting the specific subtype of EBV-positive gastric carcinoma indicates that the compound MC180295 can inhibit tumor growth by suppressing DNA repair, arresting the cell cycle, and reactivating latent EBV[51]. Additionally, a strategy involving the targeted activation of demethylases, such as TET proteins, has been explored. For example, mitoxantrone has been shown to elevate 5hmC levels and induce apoptosis in GC cells[52]. Although these compounds are still in the preclinical stage, they reveal the feasibility of developing precise drugs targeting different methylation regulatory nodes and specific GC subtypes, such as the EBV-positive type, paving the way for more targeted clinical trials in the future.

Beyond developing more specific methyltransferase inhibitors, directly using demethylating agents to overcome therapeutic resistance is another approach. Research has found that cisplatin resistance in EBV-associated gastric carcinoma is related to DNMT3A-induced hypermethylation of the ATM promoter and the subsequent silencing of DNA repair pathways. Preclinical models confirm that using the demethylating agent 5-azacytidine can restore tumor sensitivity to cisplatin by reversing ATM promoter methylation[53]. This provides an important mechanistic basis and potential strategy for using epigenetic modulators to reverse clinical chemotherapy resistance. Similarly, promoter methylation of the SLFN11 gene leads to its silencing in GC and is associated with resistance to DNA-damaging agents[46]. These findings suggest that in future clinical practice, detecting specific methylation profiles in tumor tissue may help predict patient sensitivity to platinum-based or DNA-damaging drugs, thereby guiding more individualized chemotherapy regimen selection.

Although the prospects are considerable, significant challenges remain in translating these strategies to the clinic: How to improve drug targeting to tumors to minimize the global impact on the epigenome of normal tissues; how to overcome the significant epigenetic heterogeneity of GC to ensure efficacy; and how to assess the potential risks of long-term medication. Addressing these issues is crucial for achieving clinical translation. Future research needs to focus on developing targeted delivery systems, exploring patient stratification methods based on methylation subtypes, and promoting the design of rigorous early-stage clinical trials.

Methylation marker-guided chemoprevention

Risk stratification based on DNA methylation markers provides a key tool for achieving precise chemoprevention of GPL, allowing preventive measures to target truly high-risk populations. The advantage of methylation markers lies in their ability to dynamically reflect risk changes. For example, studies show that methylation levels of CDH1 in the gastric mucosa decrease following H. pylori eradication[54]. This finding has direct clinical management implications: For patients who have completed H. pylori eradication therapy, regular monitoring of methylation level changes in key genes such as CDH1 in gastric mucosa or peripheral blood can serve as an objective molecular indicator for assessing the preventive efficacy of eradication therapy and dynamically adjusting subsequent monitoring intervals, thereby enabling more refined risk management and control.

Furthermore, research reveals that some interventions may exert preventive effects by modulating epigenetics. The work by Xu et al[55] demonstrated that postoperative supplementation with ω-3 fatty acids in elderly GC patients reduced tumor necrosis factor alpha promoter methylation in natural killer (NK) cells and enhanced NK cell activity. This provides a prospective molecular biological basis for considering ω-3 fatty acids as part of a postoperative adjuvant nutritional support regimen for GC patients, aiming to improve immune status and potentially reduce recurrence risk. Another preclinical study found that the natural polyphenol gallic acid could effectively suppress IM and dysplasia in murine models by inhibiting the Wnt/β-catenin pathway[56]. This suggests that screening and developing low-toxicity, highly effective methylation modulators from natural products is a highly promising direction in the field of chemoprevention. Future preventive clinical trials of such natural products in high-risk populations could be considered.

To achieve the clinical implementation of methylation-guided chemoprevention, a standardized pathway needs to be established in the future. First, validate and standardize methylation risk stratification panels applicable to endoscopic or liquid biopsies. Second, based on stratification results, design and implement prospective intervention studies comparing the effects of different preventive measures, such as nutritional supplementation or drugs, in high-risk subgroups. Finally, conduct health economic evaluations. This pathway can promote the transition of methylation profile-based precision prevention from theory to clinical practice.

Figure 3 illustrates the mechanisms of targeted intervention based on DNA methylation in GPL.

Figure 3 Mechanisms of targeted intervention based on DNA methylation in gastric precancerous lesions. EBV: Epstein-Barr virus.

CURRENT CHALLENGES AND FUTURE DIRECTIONS

Intratumoral epigenetic heterogeneity is a major obstacle for the clinical application of DNA methylation markers. For instance, studies in esophageal cancer reveal only 62.5% concordance in p16 methylation between endoscopic biopsies and surgical specimens, indicating that tumor heterogeneity can lead to underdetection of methylation markers in GPL and highlighting the need for optimized sampling strategies[57]. Efforts to overcome resistance include novel strategies such as CRISPR screening, which identified the histone demethylase LSD1 complex as a suppressor of EBV lytic reactivation through maintenance of H3K4 hypomethylation. Inhibition of LSD1 in GC models reactivated the virus and enhanced ganciclovir sensitivity[58].

CONCLUSION

DNA methylation, as a central epigenetic mechanism, plays a critical role in the initiation, progression, and malignant transformation of GPL. This review provides an integrated overview of the mechanistic basis and translational implications of DNA methylation in precancerous states such as chronic atrophic gastritis, IM, and dysplasia. Aberrant promoter hypermethylation leading to silencing of TSGs, including EYA4, RB1, and CDH1, consistently emerges as a key driver of lesion progression. H. pylori infection, as a primary trigger, induces epigenetic alterations through virulence factors such as CagA, chronic inflammation, and DNMT activity modulation. Additional influences such as host aging, metabolic reprogramming, and viral integration further shape the heterogeneity and clinical behavior of GPL. Clinically, DNA methylation markers detected in blood, stool, and tissue specimens—including mSEPT9, mRNF180, and SDC2—exhibit strong potential for early detection, risk stratification, and prognostic evaluation. Multiple studies indicate that multi-marker panels often provide superior sensitivity and specificity for detecting early GC and advanced precancerous lesions compared with conventional serological or endoscopic prescreening. Of particular note is the use of methylation markers in risk stratification. Specific methylation patterns have been shown to predict the risk of IM progression or the likelihood of metachronous cancer after endoscopic resection, thereby enabling more personalized surveillance intervals and interventions. Furthermore, methylation profiles are associated with prognostic outcomes, immune microenvironmental characteristics, and therapeutic responsiveness, consolidating their role in precision medicine. From a therapeutic perspective, the reversibility of methylation creates opportunities for targeted interventions. Approaches include targeting epigenetic enzymes—for instance, using DNMT inhibitors to reactivate silenced TSGs—or exploiting synthetic lethality in cancers with specific methylation defects. Chemoprevention strategies guided by methylation-based risk profiles are also emerging, with changes in genes such as CDH1 serving as biomarkers of preventive efficacy. Despite substantial progress, challenges remain, including intratumoral epigenetic heterogeneity, inconsistent marker validation, and drug resistance mechanisms. Future research should prioritize multi-omics integrative analyses, the application of spatial epigenomic-transcriptomic technologies, and the development of interventions tailored to specific epigenetic subtypes.