A hallmark of metabolic reprogramming in GC cells is the canonical Warburg effect, whereby tumor cells preferentially rely on glycolysis for ATP production even under aerobic conditions, accompanied by the suppression of mitochondrial oxidative phosphorylation. This metabolic phenotype ultimately drives robust intracellular lactate production and accumulation (18–20) (Fig. 1). In GC cells, hypoxia-inducible factor 1α and c-Myc serve as the core transcriptional regulators driving the sustained activation of the glycolytic pathway. Both factors transcriptionally upregulate key glycolytic molecules, including glucose transporter 1 (GLUT1, encoded by SLC2A1), hexokinase 2 (HK2), pyruvate kinase M2 (PKM2), and lactate dehydrogenase A (LDHA), which markedly enhance cellular glucose uptake and glycolytic flux (21). Among these, upregulated expression of GLUT1 significantly augments transmembrane glucose transport, providing sufficient substrates for glycolytic reactions (22). Intracellular glucose is first catalyzed by HK2 to generate glucose-6-phosphate, which is then converted to pyruvate via a cascade of sequential enzymatic reactions. As the terminal rate-limiting enzyme of the glycolytic pathway, PKM2 acts as a core modulator of glycolytic flux (23).

Pyruvate derived from glycolysis is catalyzed by LDHA and ultimately reduced to lactate, coupled with the oxidation of NADH to NAD+. This process completes the regeneration of coenzymes required for the glycolytic cycle, thus ensuring sustained operation of the pathway (24). Transmembrane lactate transport is mainly regulated by two members of the monocarboxylate transporter (MCT) family: MCT1 and MCT4 (25,26). MCT4 is abundantly expressed in the plasma membrane of GC cells, and effluxes excess intracellular lactate into the TME via a monocarboxylate-proton co-transport mechanism. By contrast, MCT1 primarily mediates the uptake of lactate from the TME into cells to support cellular oxidative metabolism (27–29). Together, these two transporters mediate the intracellular-extracellular lactate cycle, which ultimately leads to massive lactate accumulation and the formation of a characteristic acidic TME of GC (30).

Glutaminolysis is another hallmark of GC metabolic reprogramming (Fig. 2). Glutaminase (GLS) catalyzes the conversion of glutamine to glutamate, which is further metabolized to α-ketoglutarate for entry into the tricarboxylic acid (TCA) cycle. This process provides critical metabolic precursors that support ATP generation and macromolecule biosynthesis in tumor cells (31). GLS is abundantly expressed in GC tissues (32,33), and its dysregulation synergistically drives lactate accumulation via two distinct pathways. First, by replenishing TCA cycle intermediates, GLS reduces the mitochondrial oxidative catabolism of pyruvate, thereby shunting pyruvate toward lactate production (34). Second, glutamine is an essential nitrogen source for nucleotide biosynthesis (35). Aspartate deficiency during glutamine metabolism induces the expression of the glutamine transporter alanine-serine-cysteine transporter 2 by activating transcription factor 4 (36), thus forming a positive metabolic feedback loop that further augments the Warburg effect and promotes lactate production. Inhibition of GLS activity significantly reduces lactate production in GC cells while concomitantly suppressing tumor cell proliferation and reversing the chemoresistant phenotype (37).

In terms of fatty acid metabolism, GC cells employ a sterol regulatory element-binding protein-1 (SREBP-1)-mediated transcriptional program to activate enzymes such as fatty acid synthase (FASN) and acetyl-CoA carboxylase, converting acetyl-CoA into long-chain fatty acids for the synthesis of membrane phospholipids and signaling molecules (38). This process is regulated by the PI3K/AKT/mTOR pathway: mTORC1 phosphorylation promotes the nuclear translocation of SREBP-1 and enhances the expression of genes associated with fatty acid synthesis, including FASN, stearoyl-CoA desaturase 1, and ATP citrate lyase (39). As a transcriptional hub, SREBP-1 directly binds to the promoters of ATP citrate lyase and acyl-CoA synthetase short-chain family member 2, systematically regulating lipid metabolism at the transcriptional level (40). Concurrently, carnitine palmitoyl transferase 1A mediates mitochondrial β-oxidation of long-chain fatty acids, breaking them down into acetyl-CoA to replenish the TCA cycle and maintain energy metabolism in tumor cells (41). This metabolic flexibility between synthesis and catabolism provides membrane structural materials for GC cell proliferation and supports cell survival through energy reserves, thus representing a critical feature of metabolic reprogramming in the GC microenvironment (Fig. 3).

The synergistic remodeling of glycolysis and glutamine and fatty acid metabolism provides an energy and material basis for GC cells. Lactate, the end product of glycolysis, is not only involved in microenvironment regulation but also mediates lactylation as a new mode of epigenetic regulation, emerging as a bridge connecting metabolic reprogramming and gene expression networks.

Writer enzymes are core molecules that catalyze the conjugation of lactyl groups to lysine residues of target proteins and initiate the lactylation modification process. Notably, some of the currently identified lactylation-writer enzymes share an enzymatic system with acetylation modifications, which is an important research focus in this field.

p300, a member of the histone acetyltransferase family, was the first enzyme identified to mediate histone lactylation (11). It possesses dual catalytic activity as both an acetyltransferase and lactyltransferase, representing a key molecule for the overlapping enzymatic systems of lactylation and acetylation. This discovery once sparked academic controversy over whether lactylation is merely a concomitant byproduct of acetylation. As a pivotal writer enzyme for lactylation, p300 can specifically transfer lactyl groups to lysine residues of histones, such as histone H3 lysine (H3K9) and histone H3 lysine 18 (H3K18) (43); mediate chromatin structural remodeling by altering the binding affinity between histones and DNA; and subsequently activate the transcriptional expression of downstream target genes (44), thereby serving as a critical molecular hub connecting cellular metabolic signals to epigenetic regulation.

However, lactylation and acetylation at the same site are not functionally redundant. Instead, they compete for substrates and exhibit mutually exclusive modification sites, leading to distinct transcriptional outputs. Acetylation neutralizes the positive charge of lysine via a 2-carbon group, achieving broad-spectrum chromatin relaxation (45); By contrast, the 3-carbon group of lactylation carries a hydroxyl moiety, which not only neutralizes charge but also confers greater steric hindrance and hydrophobicity, specifically remodeling the chromatin conformation at the promoter regions of target genes (11). Both modifications also differ in their specific reader proteins: Acetylation is recognized by bromodomain containing (BRD)2 and BRD4 (46) to regulate housekeeping genes associated with cellular homeostasis, whereas lactylation is specifically bound by double PHD fingers 2 (47). In the high-lactate microenvironment of GC, p300 preferentially catalyzes lactylation (13), which precisely activates glycolytic reprogramming and drives the malignant phenotype of GC, showing a sharp divergence from the broad-spectrum transcriptional output of acetylation.

Histone acetyltransferase binding to ORC1 (HBO1), a canonical histone acetyltransferase of the MYST family, classically functions to catalyze acetylation at histone sites including histone H4 lysine 16 and histone H3 lysine 14, thereby regulating chromatin structure and gene transcription (48,49). HBO1 also exhibits novel lysine lactyltransferase activity and can specifically catalyze histone lactylation at the H3K9 site (50), endowing it with new functional implications in metabolically-dependent epigenetic regulation. Of particular importance, HBO1 is abundantly expressed in GC tissues (51), and its overexpression is negatively correlated with the 5-year survival rate of patients with GC (52), suggesting that it may contribute to the malignant progression of GC by mediating histone lactylation and exerting a non-negligible impact on patient prognosis. Using stable isotope labeling with amino acids in cell culture-based quantitative proteomics, researchers further identified 95 endogenous Kla sites regulated by HBO1, among which histone-related modification sites account for 32%, including key sites such as H3K9 lactylation and histone H4 lysine 5 lactylation. These results directly validated the core writer enzyme function of HBO1 in histone lactylation (50).

Alanyl-tRNA synthetase 1 (AARS1) is an enzyme involved in lactylation. Unlike p300 and HBO1, AARS1 does not depend on canonical lactyl-CoA as a lactyl group donor. Instead, it directly binds to lactate and catalyzes the formation of a lactate-AMP intermediate, thereby covalently transferring lactyl groups to lysine residues of target proteins to achieve lactylation (53). This discovery expands the molecular framework underlying lactylation modifications.

Eraser enzymes are core molecules that specifically remove lactyl groups from the lysine residues of target proteins and reverse the lactylation process. They maintain the dynamic homeostasis of intracellular lactylation by negatively regulating excess lactylation (54,55). Histone deacetylase (HDAC)1-3 isoforms have been validated as specific eraser enzymes for lactylation (56). They selectively recognize lactylated lysine residues and precisely remove lactyl groups, thereby mediating the targeted regulation of lactylation. In addition, sirtuin (SIRT)1 and SIRT2 have also been verified to possess lactylation erasure activity (57). In GC, these two enzymes suppress the malignant phenotype and progression of tumor cells by specifically reducing the levels of H3K18 lactylation (H3K18la) modification (58).

There is an ongoing scientific controversy with regard to the exact mechanism by which lactylation modifications occur; the debate primarily centers on whether enzyme-dependent or non-enzymatic spontaneous lactylation processes are involved (59). Enzyme-dependent lactylation is a well-validated core regulatory mechanism in GC research. This process relies on the catalysis of writer and eraser enzymes and is characterized by site specificity and regulability (59,60).

In contrast to enzymatic lactylation, non-enzymatic lactylation originates from methylglyoxal (MGO), a byproduct of glycolysis (60), and is primarily mediated by the glyoxalase pathway (61). The intracellular concentration balance of MGO is maintained by the glyoxalase cycle, which requires the coordinated action of glyoxalase 1 (GLO1) and glyoxalase 2 (GLO2). GLO1 first catalyzes the isomerization of the hemithioacetal formed by glutathione (GSH) and MGO to generate S-D-lactoylglutathione (LGSH). GLO2 then hydrolyzes LGSH to recycle GSH and produce D-lactate (62). LGSH covalently attaches lactoyl groups to protein lysine residues via non-enzymatic acyl transfer to form lactoyl-lysine modifications, establishing a link between glycolytic by-products and protein post-translational modifications, in which GLO2 plays a major regulatory role (63,64).

Although the basic regulatory mechanism of non-enzymatic lactylation has been preliminarily elucidated, there is no direct in vivo experimental evidence confirming the existence of functional non-enzymatic lactylation in GC tissues. Furthermore, it remains to be verified whether non-enzymatic and enzyme-dependent lactylation act synergistically to enhance the pro-tumor effects of lactylation, or competitively occupy modification sites to antagonize the normal regulation of enzyme-dependent lactylation during GC tumorigenesis and progression.

Increasing evidence over recent years has established that lactylation is not merely a passive byproduct of metabolic reprogramming but a critical nexus bridging cellular energy metabolism and epigenetic regulation. Lactylation modulates the activity of the glycolytic pathway at two fundamental levels, namely epigenetic transcriptional control and protein post-translational modification, thereby driving an epigenomic-metabolic cooperative regulatory circuit that represents a core mechanism sustaining the persistent activation of aerobic glycolysis in GC cells (65).

Lactylation marks have been found to be enriched in the promoter regions of key glycolytic genes, including GLUT1, HK2, and LDHA. At these loci, lactylation upregulates target gene transcription by remodeling chromatin accessibility, thereby directly activating the glycolytic cascade. Specifically, lactylation of LDHA at lysine 5 increases the catalytic efficiency of pyruvate-to-lactate conversion by 40% via an allosteric effect, while potently inhibiting the reverse reaction to maintain lactate homeostasis (66). Lactylation of HK2 at lysine 382 enhances its interaction with voltage-dependent anion channel 1, thereby directing >80% of the glucose-derived carbon flux into the glycolytic pathway (67). Furthermore, the lactylation of PKM2 at lysine 433 promotes its nuclear translocation, which in turn activates c-Myc expression via histone H3 phosphorylation. This modification leads to a 25% increase in the proportion of cells in the S phase of the cell cycle, achieving cross-functional regulation of this metabolic enzyme beyond its canonical catalytic role (68).

Clinical studies have validated the regulatory role of lactylation. Specifically, GC tissues exhibit significantly higher expression of LDHA mRNA and lactate content than paired adjacent non-tumor tissues (69). H3K18la expression was shown to be positively correlated with the expression of LDHA and HK2, and patients with concurrently high expression of both Nijmegen breakage syndrome protein 1 K388 lactylation and LDHA had a significantly poorer overall survival (70).

In addition to its core regulatory function in the glycolytic pathway, lactylation acts via alternative mechanisms to cooperatively sustain the survival and proliferation of GC cells. Lactylation of NOP2/Sun RNA methyltransferase 2 (NSUN2) at lysine 508 (K508) enhances its enzymatic activity and preserves the stability of the 5-methylcytosine (m5C) modification of glutamate-cysteine ligase catalytic subunit (GCLC) mRNA, thereby augmenting the antioxidant capacity of tumor cells (71). Under copper stress, the lactylation of methyltransferase-like 16 at lysine 229 triggers cuproptosis, a process that is potently inhibited by SIRT2. Accordingly, combined treatment with copper ionophores and SIRT2 inhibitors synergistically induces apoptosis of GC cells (72). As a bona fide lactate transferase, AARS1 senses intracellular lactate concentrations and subsequently translocates to the nucleus, where it activates the Yes-associated protein-TEA domain transcription factor complex through lactylation, and in turn forms a positive feedback loop with the Hippo pathway to continuously amplify the oncogenic effects of glycolysis (73).

The regulatory mechanisms of the aforementioned metabolic microenvironment demonstrate that histone lactylation and non-histone lactylation do not function independently. Following a sequential pattern where histone lactylation initiates metabolic reprogramming and non-histone lactylation sustains and amplifies malignant phenotypes, they form a synergistic regulatory network by sharing metabolic substrates and responding to lactate concentration gradients in a graded manner, thus serving as the core regulatory axis driving the malignant progression of GC.

Histone lactylation is a pivotal driver of lactate-mediated remodeling of the GC tumor immune microenvironment (TIME). Among the identified histone lactylation sites, H3K18la is the most extensively characterized to date, and its pro-tumor and immunomodulatory effects are mediated primarily via two major mechanisms. H3K18la directly targets and activates the transcription of vascular cell adhesion molecule 1, thereby enhancing the intrinsic malignant phenotype of GC cells via the AKT/mTOR signaling pathway. Second, H3K18la upregulates the expression of the chemokine C-X-C motif chemokine ligand 1 to recruit GC-associated mesenchymal stem cells and M2-polarized tumor-associated macrophages to the TME, which further amplifies the immunosuppressive effect (74). Furthermore, as a major source of lactate in the TME, cancer-associated fibroblasts secrete high levels of lactate that specifically induce H3K18la in GC cells, thereby regulating and activating the expression of its downstream target gene assembly factor for spindle microtubules (ASPM). ASPM directly binds to the non-SMC condensin 1 complex subunit G (NCAPG) protein to promote its nucleocytoplasmic translocation and enhances the BUB3-mediated deubiquitination of NCAPG, which markedly increases the stability and expression of NCAPG and ultimately mediates the immune evasion of GC cells (75).

At the regulatory circuit level, the histone delactylase SIRT1 negatively regulates the transcription of long noncoding RNA H19 by specifically erasing H3K18la marks. In turn, H19 forms a positive feedback regulatory loop with LDHA, the key rate-limiting enzyme for lactate production, and H3K18la, which sustainably amplifies the pro-tumor effects and immune microenvironment remodeling mediated by histone lactylation (58). Single-cell RNA sequencing analysis of human gastric cardia adenocarcinoma tissues has revealed that elevated H3K18la modification levels are associated with impaired CD8+ T-cell function (76), providing clinical evidence supporting the critical role of histone lactylation in GC progression and TIME regulation. By contrast, the specific biological roles and underlying molecular mechanisms of non-histone lactylation in the remodeling of the GC immune microenvironment remain poorly defined and warrant further in-depth investigation (Fig. 5).

Lactylation influences tumor cell proliferation, invasion, and immune escape by connecting energy metabolism and epigenetic regulation. However, the mechanisms by which lactylation-related genes integrate metabolic features, drug responsiveness, and immune infiltration patterns remain to be elucidated. Recent studies on the lactylation regulatory network in GC have provided newer insights into disease prediction and therapy (Table I).

Zhang et al (77) developed a predictive model based on ‘hypoxia-glycolysis-lactylation-related genes (HGLRG)’. By integrating bulk and single-cell RNA sequencing data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus databases, they identified a prognostic signature comprising 13 genes, including CP, biglycan (BGN), DUSP1, SERPINE1, and CLDN9. Functional correlation analysis has revealed that BGN expression was significantly associated with amino acid and lipid metabolites, suggesting a critical role in metabolic reprogramming. Immunophenotypic analysis has further indicated that the HGLRG signature was correlated with M1 macrophage and CD8+ T-cell infiltration, providing a theoretical basis for combined therapeutic strategies targeting metabolic pathways and the immune microenvironment (77).

Yin et al (78) analyzed lactylation-related and mitochondrial function genes in stomach adenocarcinoma based on multi-omics data and observed significant prothymosin α (PTMA) overexpression in cancer tissues. Knockdown of PTMA inhibited gastric cell proliferation, migration, and invasion while inducing apoptosis, revealing its oncogenic role in regulating mitochondrial function and metabolic reprogramming (78).

Yang et al (79) screened six core lactylation-related genes from a Gene Set Enrichment Analysis database and developed a GC lactylation scoring model. In a multi-cohort analysis, they observed that the lactylation score was correlated with the overall survival, tumor progression, and response to immunotherapy of the patient. High-scoring populations exhibited a higher risk of immune escape. Mechanistically, LDHA inhibitors suppress malignant tumor phenotypes by downregulating the expression of lactylation-related genes such as prolyl 3-hydroxylase domain-containing protein 2 (PLOD2) and glucose transporter type 3 (GLUT3) (79). Clinical studies have further substantiated that an increased expression of PLOD2 and GLUT3 is closely associated with tumor invasion, metastasis, and poor prognosis (80,81).

The prevailing TCGA molecular classification system for GC is primarily based on genomic and epigenomic features. By contrast, the aforementioned lactylation-related gene signature and its corresponding scoring model have enabled an integrated characterization of cancer metabolic reprogramming, heterogeneity of the TIME, and therapeutic response profiles. Thus, this model addresses critical unmet gaps in the TCGA classification system with respect to the prediction of metabolic phenotypes, the immunosuppressive TME, and immunotherapeutic response, and ultimately provides a more robust and precise stratification framework for the genomically stable (GS) and chromosomal unstability (CIN) subtypes of GC, which are characterized by pronounced prognostic heterogeneity.

The lactylation-related gene signatures and scoring models can further integrate information on tumor metabolic reprogramming, immune microenvironment heterogeneity, and therapeutic responses. They complement the predictive blind spots of the GC molecular classification system of TCGA with regard to metabolic phenotypes, immunosuppressive microenvironments, and immunotherapy response, and can provide more precise stratification criteria for GS- and CIN-type GC, which display marked prognostic heterogeneity.

Lactate accumulation in the TME drives metabolic reprogramming and epigenetic remodeling in GC by promoting histone and non-histone lactylation. Interventions targeting the upstream and downstream pathways of lactylation demonstrate antitumor potential through metabolic regulation and immune microenvironment remodeling.

In the acidic TME, N-α-acetyltransferase 10 (NAA10)-mediated K508 lactylation of NSUN2 activates its enzymatic activity, promoting m5C methylation and the stability of GCLC mRNA, enhancing glutathione synthesis to inhibit lipid peroxidation, and inducing doxorubicin resistance. Targeted blockade of NSUN2 lactylation disrupts drug resistance, restores sensitivity to ferroptosis-inducing drugs, and enhances chemotherapeutic efficacy (71).

The cancer cell membrane-camouflaged nanodelivery system IRCB@M was developed based on the glutaminolysis inhibitor CB-839 to allow co-delivery of CB-839 and photosensitizer IR-780 via homologous targeting. CB-839 suppresses the glutamine-dependent antioxidant defense system and reduces nicotinamide adenine dinucleotide phosphate and GSH levels, while markedly enhancing the apoptotic effect of photodynamic therapy in tumor cells, thus achieving coordinated regulation of tumor energy metabolism and redox homeostasis (82).

Regarding the remodeling of the TIME and metastasis inhibition, MCT1 acts as a key target for reversing immunosuppression in GC. The MCT1 inhibitor AZD3965 not only suppresses the abnormal expression of the pro-fibrotic factor fibroblast activation protein (FAP) in mesenchymal stromal cells, which attenuates stromal fibrosis and tumor-promoting proliferation, but also downregulates the expression of programmed death-ligand 1 (PD-L1) in tumor cells and the metastasis-related gene MACC1. It simultaneously relieves immune evasion and blocks invasion and metastasis, thereby restraining the malignant progression of GC via dual pathways (83).