Gastric cancer (GC) is one of the most common malignancies and the third leading cause of cancer-related mortality worldwide (1–3). The etiology of GC is multifactorial, and is influenced by genetic alterations, dietary practices, lifestyle factors, atrophic gastritis, Helicobacter pylori and Epstein-Barr virus (EBV) infections (4,5). Notably, >70% of patients with GC are currently diagnosed at an advanced stage due to the absence of routine screening methods and the asymptomatic nature of early disease. Additionally, this group of patients constitutes a population with a high mortality rate, due to the lack of targeted therapies, resistance to chemoradiotherapy and distant metastases. Therapeutic success rate is high for patients with early-stage gastric cancer, with 5-year survival rates of >90% (6); among patients with advanced-stage disease, the prognosis remains poor, with a 5-year survival rate of 5–10% (7). Therefore, novel effective therapeutic approaches are urgently needed to improve the clinical outcomes of patients with GC.

Immunotherapy has transformed GC research from a purely tumor-centered model into a broader immune-biology framework, greatly advancing the understanding of its molecular characteristics, tumor heterogeneity, immune interactions, molecular subtypes, tumor immune microenvironment, immune escape mechanisms and biomarker discovery, while demonstrating considerable therapeutic potential. Multiple immunotherapies have been designed to activate the immune system and enhance antitumor responses in patients with GC, including immune checkpoint inhibitors (ICIs) [such as nivolumab (8) and pembrolizumab (9)] and antigen-targeted chimeric antigen receptor (CAR)-T cell therapies, which include claudin 18.2 (CLDN18.2), mucin 3A, mesothelin (MSLN) and carcinoembryonic antigen (CEA). Notably, CLDN18.2-targeted therapies, such as the monoclonal antibody zolbetuximab, have demonstrated survival benefits in phase III trials (10,11), and CLDN18.2-targeted CAR-T-cell therapies have shown early but promising efficacy (12). However, the efficacy is notably limited by the immune cell composition of the tumor microenvironment (TME). As chronic inflammation is a critical component in the development of human GC, the TME in GC presents a highly dynamic and complex ecosystem that comprises diverse extracellular matrix proteins, secretory proteins and cellular components. Cellular components include cancer-associated fibroblasts (CAFs), endothelial cells and immunosuppressive cell populations, including regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs) and tumor-associated macrophages (TAMs) (13–15). The TME in GC has also been shown to be rich in pro-inflammatory cytokines, chemokines and growth factors; the complex interaction between these factors and receptors influences the growth and progression of GC (16).

Chemokines and their receptors are widely expressed in tumor cells, immune cells and stromal cells within the TME. Chemokines are pivotal in mediating directional cell migration, particularly in leukocytes (for example, CXCL8-CXCR1/CXCR2, CXCL9/10/11-CXCR3, CXCL16-CXCR6 and CX3CL1-CX3CR1) (17–20) and promoting angiogenesis (such as CXCL5/CXCL6-CXCR2 and CXCL12-CXCR7) (21,22). Moreover, some chemokines contribute to carcinogenesis by modulating tumor transformation (for example, CCL19/21-CCR7 and CCL2-CCR2) (23,24), invasion (such as CXCL5-CXCR2) (25) and metastasis (such as CXCL8-CXCR1/2 and CXCL12-CXCR4) (18,26). In GC, a comprehensive understanding of the complex interconnected networks of secretory factors within the TME may help to identify new targets for diagnostic and prognostic assessments, and therapy.

The present review presents a comprehensive analysis of the TME in GC, providing novel insights into the cell-chemokine interaction networks in GC and highlighting the therapeutic implications of chemokines in immunotherapy. In addition, current related research areas and therapeutic advances associated with key chemokine signaling pathways in GC are summarized. Specific attention has been paid to the progress and ongoing challenges of chemokine-targeted CAR-T immunotherapy in modulating the immune landscape of the TME in GC. By providing critical insights into TME reprogramming and immunomodulation, the current review aims to advance the strategic application of chemokine-targeted CAR-T immunotherapy in GC.

The TME in GC is a highly dynamic and complex ecosystem (27,28). Single-cell sequencing analyses have revealed that the TME in GC comprises various immune cells, including dendritic cells (DCs), T cells and B cells, as well as cellular components, including epithelial cells (ECs), MDSCs, TAMs, tumor-associated neutrophils and CAFs (29). Tumor cells within the TME secrete a range of soluble factors, including chemokines, cytokines, growth factors and extracellular matrix components; these factors contribute to the establishment of tumor-protective barriers and promote tumor growth, angiogenesis and metastasis (30,31). Similar to the TME in a number of other solid tumors, the TME in GC is immunosuppressive, promoting tumor progression and the emergence of an aggressive phenotype. In the GC TME, cells interact through cytokines and chemokines, as shown in Fig. 1. The immune cells within this TME interact primarily through the secretion of cytokines or chemokines that can be broadly categorized as antitumorigenic [such as IL-12, interferon (IFN)-γ, TNF-α, TNF-related apoptosis-inducing ligand, CCL1, CCL19, CCL21, CXCL10 and CXCL16] or pro-tumorigenic (for example, IL-6, IL-23, IL-10, IL-17, TNF-β, MMP-9, CCL2, CCL5, CCL7, CCL12, CCL17, CCL22, CXCL8, CXCL9, CXCL10 and CXCL12) (32–35).

There are two distinct types of macrophages based on cytokine expression, including M1 and M2. M1 macrophages, activated by IFN-γ and microbial components such as lipopolysaccharide, exhibit potent tumor-killing capacity; this type of macrophage expresses high levels of major histocompatibility complex (MHC) class II molecules and produces pro-inflammatory cytokines, including IFN-γ and TNF-α (36). Conversely, M2 macrophages, which are activated by IL-4, produce immunosuppressive cytokines (such as IL-10), and exhibit poor antigen-presenting capacity and tumoricidal activity (37). Under certain conditions, macrophages are recruited to the TME and polarize into TAMs (M1 or M2). Most TAMs present as M2 macrophages that lack the ability to phagocytose tumor cells and serve a crucial role in tumor initiation, invasion, metastasis and immune evasion (38,39). Furthermore, TAMs are critical for modulating the programmed cell death protein 1 (PD-1)/programmed death-ligand 1 axis and promoting the progression of GC cells, thereby contributing to a poor prognosis in patients.

Based on the immunosuppressive role of M2-type TAM in tumor progression and immune evasion, several antitumor strategies aimed at depleting M2-type TAMs or reprogramming them to M1-type have been developed. One study reported an immunosuppressive M2-like profile in a subset of TAMs expressing folate receptor β (FRβ); this study revealed that preconditioning the TME with FRβ-specific CAR-T cells improved the efficacy of tumor-directed anti-MSLN CAR-T cells (40). Furthermore, Zhang et al (41) successfully reversed the immunosuppressive, tumor-supportive state of TAMs by delivering mRNA encoding transcription factors IFN regulatory factor 5 and IKKβ. This strategy could reprogram M2-like TAMs into an M1-like phenotype, inducing antitumor immunity and promoting tumor regression. Overall, promoting polarization to the M1 phenotype within the TME offers a promising therapeutic strategy.

CAFs are a heterogeneous population of activated fibroblasts and a central element of the TME; this population is pivotal in tumor progression and metastasis by promoting the migration of cancer cells, altering the metabolism of epithelial tumor cells, regulating the metabolic flexibility of cancer cells and contributing to the development of therapeutic resistance (42). Moreover, CAFs, are a key constituent of the TME that favor cancer progression via the secretion of molecules, including cytokines (such as IL-6, IL-11 and TGF-β), chemokines (for example, CCL2, CXCL8, CXCL9, CXCL10 and CXCL12) and growth factors (for example, fibroblast growth factor). These molecules can directly activate GC cells and promote aggressive phenotypes. It has been reported that a specific subset of CAFs associated with poor prognosis can recruit macrophages into the TME via the C3-C3AR1 axis, and this interaction may govern the response to immune checkpoint blockade in GC peritoneal metastases (GCPM). In addition, chemokines, such as CCL2, CCL3, CCL4 and CXCL8, have been demonstrated to exhibit increased expression during GCPM progression; these chemokines facilitate the establishment of an immunosuppressive microenvironment by recruiting inflammatory monocytes, neutrophils and TAM-like macrophages into the peritoneum. The recruited TAM-like macrophages secrete TGF-β, which induces the conversion of peritoneal mesothelial cells into CAFs, thus promoting the formation of pre-metastatic niche (43). Notably, CCR2-positive GC cell subsets have been observed to possess the ability to directly sense the CCL2 gradient on the peritoneal surface and migrate directionally toward higher chemokine concentrations, achieving directed colonization (44). Research has also reported an elevation of pro-inflammatory cytokines (such as IL-6 and TNF-α) in the TME, which can activate the STAT3 pathway in GC cells, and further enhance cell adhesion, survival and implantation on the peritoneal surface (45).

Another study showed that CAFs can suppress the proliferation and migration of CD8⁺ cytotoxic T cells into tumors partly driven by the secretion of immunomodulatory factors, such as TGF-β and chemokines (46). Moreover, CAFs have been observed to attract, support the accumulation and promote the survival of FOXP3⁺ Tregs in various types of cancer (47). It has also been reported that CAFs recruit MDSCs into the TME via CCL2 release, thereby suppressing CD8⁺ T-cell proliferation and IFN-γ production (48). These coordinated effects establish a profoundly immunosuppressive TME.

Fibroblast activation proteins (FAPs) are membrane-bound serine post-prolyl peptidases with endopeptidase activity, which are expressed on CAF subsets in various tumors and represent a promising therapeutic target. Fang et al (49) developed a monoclonal antibody linked to a tubulin-binding maytansinoid and a bispecific antibody that could target FAP on CAFs and death receptor 5 simultaneously, which demonstrated potent antitumor activity. Another study revealed that FAP-targeting CAR-T cells can enhance the efficacy of CLDN18.2-targeting CAR-T therapy by remodeling the TME (50). Collectively, CAFs are a critical barrier to antitumor immunity.

Lymphocytes, including T and B cells originating from bone marrow progenitors, serve a pivotal role in the immune landscape of GC. They can exert antitumor immune responses and contribute to immune tolerance within the TME, a functional outcome influenced by their specific lymphocyte subset and the signals they receive (51). T cells are primarily categorized into two types based on their T-cell receptor (TCR) type: αβ-T cells and γδ-T cells. The αβ-T-cell population can be subdivided into CD8⁺ cytotoxic T cells and CD4⁺ helper T cells. The former can recognize antigens presented by MHC class I molecules, whereas the latter can interact with antigens presented by MHC class II molecules (52).

Tregs in the TME are pivotal in maintaining systemic immune tolerance. In healthy individuals, Tregs in the thymus migrate to peripheral tissues, and actively regulate the activation and proliferation of potentially auto-reactive T cells. This modulation is essential for maintaining immune homeostasis and preventing autoimmune diseases (16).

Dysfunctional T cells and Tregs in the TME are among the most extensively studied immune cell populations. Research has shown that a subset of these cells co-expresses the inducible T-cell co-stimulator and IL-1 receptor 1, which dominate the immunosuppressive process in the TME (53). In addition, expression of immuno-inhibitory molecules, such as PD-1 and cytotoxic T-lymphocyte-associated antigen 4, which are the key targets of various immunotherapies, have been observed on these cells. It has been reported that tumor cells also contribute to immunosuppression by secreting anti-inflammatory cytokines, including IL-10 and TGF-β; the increase in anti-inflammatory cytokines promotes the expansion of Tregs and the production of FAS ligands, in turn inducing the apoptosis of activated T cells (54,55).

Notably, an association between intratumoral CD8⁺ T-cell infiltration and survival has been detected in patients with GC. A meta-analysis of GC demonstrated that an increase in both intratumoral CD8⁺ and CD4⁺ T-cell infiltration is associated with improved overall survival in patients (56). Another meta-analysis has revealed the possibility of immunosuppression and worsening of GC prognosis in the presence of Tregs in the TME. However, no conclusive association between FOXP3⁺ Treg infiltration and patient outcomes has been observed (57). Based on current solid tumor research, CD8⁺ T cells are critical for achieving successful treatment outcomes, and it is essential to advocate strategies that can enhance CD8⁺ T-cell infiltration and promote their persistence in the TME.

Chemokines are a large family of structurally related cytokines, with >40 individual members and their receptors found in humans. Notably, chemokines are a class of small signaling proteins that are important for embryogenesis, hematopoiesis, mitogenicity, and innate and adaptive immunity. Based on the number and location of N-terminal cysteine residues, chemokines are divided into four families: CXC, CC, XC and CX3C (58,59). CCL1-28, a member of the CCL family, exerts chemotactic effects on monocytes, macrophages, lymphocytes and granulocytes. CXCL1-16, which belongs to the CXCL family, is associated with cell differentiation, migration and tumor growth (60). CXC chemokines containing a glutamic acid-leucine-arginine motif (ELR⁺) at the C-terminus are considered pro-angiogenic (such as CXCL6 and CXCL8); conversely, those without the ELR motif are anti-angiogenic (for example, CXCL4/9/10) (61). XCL1 and XCL2 belong to the XCL family, and are both involved in the modulation of various immune cells. Notably, XCL1 is primarily expressed in activated T cells, natural killer (NK) cells and NKT cells, whereas XCL2 is predominantly expressed in activated T cells and DCs (62). CX3CL1 from the CX3CL family mediates the adhesion and chemotaxis of mature T lymphocytes, macrophages and DCs by binding to the receptor CX3CR1 (63).

Chemokine receptor-ligand interactions are important for mediating endogenous immune cell trafficking and have been considered as a therapeutic target in immunotherapy for certain types of cancer. Clinical studies of chemokine-based therapies for malignant tumors are currently underway (Table I), highlighting chemokines as promising therapeutic targets. In addition, CAR-T cells engineered to express chemokine receptors have been proven to possess enhanced tumor trafficking and homing abilities. Animal experiments have shown that the co-expression of chemokine receptors (CCR2b, CCR8, CXCR2 and CXCR6) in CAR-T cells targeting GD2, MSLN or B7H3 enhances tumor infiltration and antitumor effects (64–67). In the current review, chemokines that are markedly expressed in GC are summarized. Furthermore, their roles in the TME and related immunotherapies are described, with the aim of providing novel insights into GC immunotherapy.

CAR-T-cell therapy has demonstrated considerable efficacy in eliminating GC cells; however, its application in solid tumors remains limited by the immunosuppression and immune evasion within the TME. Chemokine-based strategies, by which the immune profile within the TME can be regulated, hold notable potential for enhancing the efficacy of immunotherapy (Fig. 2).

Mechanistically, disruption of chemokine signaling can enhance T-cell trafficking. For example, CRISPR-Cas9-mediated knockout of chemokine receptors (such as CXCR4) on T cells (119) or use of CXCR4 inhibitors has been shown to release the T cells retained within lymphatic vessels, vascular endothelial cells or CAFs (105). Since certain tumor environments inhibit the expression of homing-related chemokine receptors, such as CXCR6, upregulating these receptors represents a promising therapeutic approach. Research has also demonstrated that engineering CAR-T cells to express chemokines such as CCL19, CCL21 and CXCL10 can increase the recruitment of DCs and NK cells, promoting antigen presentation and cytotoxic activity, and thereby enhancing antitumor immunity (74,120,121).

GC is characterized by marked heterogeneity. Distinct subtypes arise through different oncogenic mechanisms, resulting in substantial differences in TME features and chemokine expression profiles. According to The Cancer Genome Atlas classification, GC can be classified into four molecular subtypes: EBV-associated, microsatellite instability (MSI)-high, genomically stable (GS) and chromosomal instability (CIN). The EBV and MSI subtypes are usually linked to an immune-inflamed phenotype; these subtypes are characterized by increased T-cell infiltration and elevated expression of CXCL9/10/11 and CCL5, which are associated with improved responses to immunotherapy. By contrast, GS tumors exhibit an immunologically ‘cold’ phenotype, with enriched CAFs and MDSCs. Elevated levels of CXCL8, CXCL12 and CCL2 promote the recruitment of immunosuppressive cells and limit T-cell infiltration, suggesting that the CXCL8-CXCR2 axis may serve as a potential therapeutic target. CIN tumors display a fibroblast-rich and immunosuppressive microenvironment, in which CXCL8 and CXCL12 expression shows an upward trend and exacerbate immune suppression (122,123). Targeting pathways such as CXCR4 while performing CAR-T therapies, therefore, may represent an effective combinatorial strategy (124–126).

Despite these advances, the clinical safety of chemokine/chemokine receptor engineering combined with CAR-T therapy warrants careful evaluation. Notably, a marked expansion of CXCR6+ cytotoxic T cells has been detected in the cerebrospinal fluid of patients with hematological malignancies, including diffuse large B-cell lymphoma, acute lymphoblastic leukemia, mantle cell lymphoma, follicular lymphoma, highly aggressive B-cell lymphoma and multiple myeloma, who developed immune effector cell-associated neurotoxicity syndrome (ICANS) following CAR-T therapy. Moreover, CXCL16 expression in myeloid cells has been shown to be positively associated with ICANS severity. These results suggest that the CXCL16-CXCR6 axis may contribute to neuroinflammation and the pathogenesis of ICANS (127). Furthermore, in a clinical trial where CCL19- and IL-7-armored CAR-T cells were applied, the overall safety was manageable but a subset of patients developed grade 3 cytokine release syndrome and neurotoxicity (128).

To mitigate these risks, current strategies include the incorporation of inducible safety switches, and the use of mRNA co-transfection to achieve transient expression of chemokines and their receptors (129,130). Careful optimization of chemokine-engineered CAR-T-cell strategies is essential to balance the efficacy and safety in treatment of GC.

Overall, combining CAR-T-cell therapy with chemokine engineering has the potential to effectively promote CAR-T-cell accumulation and enhance their cytotoxic response within tumors. CAR-T-cell therapies incorporating chemokine engineering strategies are currently under active investigation and have demonstrated promising efficacy (Table II). These tumor-targeting studies provide valuable insights for further development of CAR-T-cell therapy for GC. CAR-T cells designed against established GC targets, such as CLDN18.2, HER2, CEA, MSLN and NKG2D, in combination with chemokines, monoclonal antibodies, bispecific antibodies and ICIs, have shown promising therapeutic potential. For example, ICIs such as nivolumab, combined with anti-HER2 antibody (trastuzumab) and chemotherapy, have demonstrated clinical benefit in patients with HER2-positive GC (131,132). Dual-target CAR-T cells incorporating chemokine signaling, and the combination of immunotherapy and radiotherapy/chemotherapy have also been observed to overcome the limitations of ‘immune-cold’ tumors and to represent a key direction for future GC immunotherapy (133,134).

The TME in GC is a highly complex and dynamic ecosystem composed of diverse cellular components, including tumor cells, CAFs, immune cells and endothelial cells. These components interact extensively to regulate tumor progression, immune evasion and therapeutic responses. Additionally, these cellular components not only shape the structural and functional characteristics of the TME but also actively influence immune surveillance and resistance mechanisms.

Chemokines serve a central role in orchestrating cellular interactions within the TME by regulating immune cell trafficking, spatial distribution and functional polarization. For example, chemokines, such as CCL19, CCL21 and CXCL16, are involved in the recruitment and organization of immune cells, thereby affecting antitumor immune responses. Dysregulation of chemokine signaling contributes to the establishment of an immunosuppressive microenvironment by recruiting Tregs, MDSCs and TAMs; meanwhile, it simultaneously inhibits the infiltration and effector function of cytotoxic T lymphocytes. For example, chemokines CXCL8 and CXCL12 are commonly associated with immunosuppression and tumor progression, whereas chemokines CXCL9 and CXCL10 are context-dependent with either pro- or antitumor effects depending on the tumor type. Accordingly, targeting chemokine-receptor axes (such as CCR2, CCR4, CCR5, CXCR2, CXCR4) has emerged as a promising strategy to remodel the TME, enhance antitumor immunity and improve therapeutic efficacy in cancers.

Despite substantial advances, the efficacy of CAR-T therapy in solid tumors such as GC remains constrained by poor tumor infiltration, antigen heterogeneity and an immunosuppressive microenvironment. Integrating chemokine-based strategies into CAR-T design, such as engineering chemokine receptors, secreting chemokines or combining with chemokine inhibitors, provides a rational approach to overcome these barriers. These strategies can enhance T-cell trafficking, promote immune cell recruitment, convert ‘cold’ tumors into ‘hot’ tumors and foster immunologically active environments.

Chemokine-based combination strategies, including dual-target CAR-T-cell therapies, mRNA co-transfection approaches, ICIs, DC vaccine, oncolytic virus-based immunology and conventional therapies (such as radiotherapy and chemotherapy), hold great promise for improving therapeutic outcomes in patients with GC. However, a number of challenges remain, particularly regarding safety, off-target effects and the lack of physiologically relevant humanized models for pre-clinical evaluation. Future efforts should focus on optimizing engineering strategies, refining combination regimens and developing advanced model systems to better predict clinical efficacy.

Overall, an improved understanding of the interplay between cellular components and chemokine networks within the TME in GC is essential for developing more effective and personalized immunotherapeutic strategies. This may contribute to improving clinical outcomes for patients with GC.