Abstract
Uncontrolled proliferation is not the only factor driving tumor growth; the tumor microenvironment (TME) has a significant impact. Tissue homeostasis and pathogen removal depend on macrophages, which are important innate immune effector cells. Nevertheless, there is growing evidence that tumor-associated macrophages (TAMs) do not consistently suppress cancer and are functionally diverse. Specifically, M2-polarized TAMs (M2-TAMs) build up in multiple solid tumors, where they promote angiogenesis, metastasis, and immunosuppression, hastening the course of the disease. Here, we critically assess the clinical translatability of TAM-targeted treatment approaches, outline the molecular circuits underpinning M2-TAMs–tumor cell interaction, and extensively explore the phenotypic spectrum and functional diversity of macrophages in cancer. Our objective is to offer a theoretical foundation for upcoming immunotherapeutic interventions.
Keywords
1. Introduction
With 5% of all deaths, cancer is the second most common cause of mortality globally and a complex illness. Neoplastic cells spread to distant tissues, have an infinite capacity for replication, and defy immune surveillance. There is growing evidence that all stages of cancer are accelerated by the heterogeneity of the TME. The TME is a dynamic ecosystem that coevolves with cancer cells and, in the end, creates a niche that specifically promotes tumor growth. 1 Immune effectors, fibroblasts, and extracellular matrix (ECM) proteins serve as gatekeepers in normal tissue; in the TME, however, these same components are appropriated to promote the advancement of disease. As a result, TME-targeting treatments have become a promising therapeutic option. 2 Macrophages play a crucial role in host defense by carrying out antigen presentation and phagocytosis to start anti-tumor immunity. The most prevalent immunological population in the majority of solid tumors, TAMs, paradoxically display pro-tumoral characteristics such as angiogenesis, metastasis, chemo-resistance, and inhibition of cytotoxic immune responses.3,4
Furthermore, exosomes serve as “messengers” to facilitate communication between cancer cells and ATMs. Tumor cell invasion, proliferation, immune evasion, and resistance to treatment are all coordinated by their cargo of proteins, miRNAs, lncRNAs, and circRNAs.5-7 However, the molecular mechanisms underlying this vesicular crosstalk are still unclear. In addition to providing theoretical support for the development of novel therapeutic approaches that block exosome function, reverse immune suppression, inhibit metastasis, and overcome drug resistance, a detailed analysis of exosome-mediated communication between macrophages and cancer cells will shed light on the regulatory topology of the tumor microenvironment.
This review, which is arranged by tumor entity, methodically outlines the reciprocal interactions between neoplastic cells and TAMs, describes the important signaling molecules and transduction cascades involved, and evaluates TAM-directed drugs that have been approved, are being developed, or are undergoing clinical testing. Our goal is to present a comprehensive map of how TAMs influence tumor growth and to pinpoint practical tactics that improve anti-cancer efficacy.
2. Exosomes Generated From Tumor Cells Coordinate Macrophages M2 Polarization, Hastening the Growth of Tumors
2.1. miRNAs Supplied by Exosomes Control Tumor-Associated Macrophages M2 Polarization
Small non-coding RNAs known as microRNAs (miRNAs) coordinate post-transcriptional gene regulation. Nearly all human malignancies have been linked to dysregulation of miRNAs in their start and development. 8 Breast cancer, the leading cause of female morbidity and mortality worldwide. Under hypoxia, neoplastic cells release exosomes enriched in miR-143-3p, whose down-regulation within these vesicles increases RICTOR expression in recipient macrophages, driving M2 polarization and enhancing tumor aggressiveness. 9 Similarly, there is a favorable correlation between M2-TAM invasion and elevated miR-205 levels in ovarian cancer specimens. By suppressing PTEN, a negative regulator of PI3K, exosome-mediated transfer of miR-205 from cancer cells to TAMs initiates the PI3K/AKT/mTOR cascade, which reprograms macrophages toward a pro-tumor phenotype. 10 As a result, miR-205 is a manageable therapeutic target for ovarian cancer. The long non-coding RNA LINC00273 is transcriptionally up-regulated by active STAT3 in M2-TAMs, which then packages it into exosomes and returns it to LUAD cells. LINC00273 binds NEDD4, inhibits the Hippo cascade, and promotes YAP’s nuclear translocation within the neoplastic compartment. After that, YAP/TEAD complexes trigger RBMX, which physically binds to pri-miR-19b-3p and encourages its selective loading into exosomes, completing a self-amplifying loop that continuously adds miR-19b-3p to the microenvironment. 11
Haematogenous dissemination is the predominant route of distant cancer spread; because the majority of gastrointestinal veins drain into the portal circulation, the liver represents the most frequent metastatic site for gastric cancer. 12 According to Qiu et al, patients with gastric cancer liver metastases(GC-LM) have significantly higher levels of miR-519a-3p in their blood exosomes. Kupffer cells internalize these GC-derived vesicles, which aggregate specifically in the liver and transmit miR-519a-3p to quiet DUSP2. Hepatic colonization is accelerated by the subsequent unchecked MAPK/ERK signaling that polarizes macrophages toward the M2 phenotype, promoting angiogenesis and creating a pre-metastatic environment. (Figure 1A) 13
2.2. Exosomal lncRNA/circRNA Cargo Drives M2 Polarisation of Tumour-Associated Macrophages
Beyond microRNAs, lncRNAs/circRNAs are emerging as critical orchestrators of malignancy. In breast cancer, intracellular lncRNA HAGLROS is frequently up-regulated and correlates with poor prognosis; its overexpression accelerates proliferation, metastasis, EMT and angiogenesis. Tumour-released exosomes enriched in HAGLROS further elevate p-STAT3 in recipient TAMs, skewing them toward the M2 phenotype and fuelling disease progression. 14 Similarly, circNEIL3 is overexpressed in glioma tissue, where it is cyclised by EWSR1 and interacts with IGF2BP3 to prevent its degradation, thereby driving proliferation, invasion and metastasis. Exosomal transfer of circNEIL3 to TAMs sustains IGF2BP3 stability and up-regulates YAP1; the resultant IGF2BP3/YAP1 axis cooperatively polarises macrophages to an M2 state, establishing an immunosuppressive micro-environment. (Figure 1B) 15
2.3. Exosomal Delivery of Signature Cargo Proteins Orchestrates the M2 Polarisation of Tumour-Associated Macrophages
In addition to RNA species, proteins—constituting a major fraction of exosomal cargo—are integral to tumour progression. Han et al reported that ovarian carcinoma cells secrete exosomes enriched in CMTM4, which trigger M2 polarisation of TAMs. These M2-TAMs subsequently activate the NF-κB pathway, releasing TGF-β1 and CXCL12. TGF-β1 enhances cancer-cell motility and invasiveness, whereas CXCL12 recruits additional monocytes/macrophages and amplifies M2 polarisation, establishing a self-reinforcing loop that accelerates ovarian cancer progression.
16
Pancreatic ductal adenocarcinoma (PDAC), one of the most lethal malignancies worldwide, carries a five-year survival rate below 10%.
17
Ma and colleagues provided actionable insight into this therapeutic roadblock: PDAC exosomes ferry CCT6A into macrophages, (Figure 1C) ignite intracellular PI3K/AKT signalling and up-regulate a spectrum of chemokines that sculpt an M2-permissive micro-environment, tilting immunity toward tumour support. Critically, high-CCT6A exosomes markedly compromise anti-CD47 immunotherapy in vivo by fostering immune tolerance.
18
Overall, exosomes can deliver various substances to tumor cells to promote tumor progression. (Table 1) Schematic illustration of the signalling crosstalk between tumour cells and TAMs. (A) LD-derived exosomes deliver miR-19b-3p to TAMs, thereby driving M2 polarisation and accelerating tumour progression. (B) Glioma-released exosomes harbour circNEIL3, which induces M2-TAM polarisation and fosters an immunosuppressive microenvironment. (C) PDAC-secreted exosomes transfer CCT6A to TAMs, promoting M2 polarisation and tumour progression Types of Exosomes Secreted by Tumor Cells
3. Macrophage-Derived Exosomes Fuel Tumour Progression by Directly Reprogramming Cancer-Cell Behaviour
Macrophages within the tumour microenvironment operate as functionally plastic and context-dependent units that exert dichotomous effects on neoplastic progression,19,20 Classically activated M1-TAMs display anti-tumour activity,21,22 whereas alternatively activated M2-TAMs predominantly foster tumour growth, invasion and immune evasion.23,24 The conventional paradigm posits that cancer cells reprogramme TAMs toward an M2-TAMs via exosomal cues, thereby sculpting a tumour-permissive niche. Emerging data, however, reveal that this intercellular communication is bidirectional: TAMs are not merely passive recipients but also active donors of exosomes that deliver functional biomolecules to cancer cells, markedly enhancing proliferative capacity, migratory and invasive potential, and, in selected contexts, inducing resistance to cytotoxic or targeted therapies.
3.1. M2-TAMs Derived Exosomes Potentiate Tumour-Cell Glycolysis, Thereby Accelerating Neoplastic Progression
Wang et al demonstrated that exosomes released by M2-TAMs are efficiently internalised by gastric cancer (GC) cells and deliver the long non-coding RNA MALAT1. Within recipient cells, MALAT1 simultaneously activates two complementary oncogenic circuits: it interacts with δ-catenin, shields it from β-TRCP-mediated ubiquitination and thereby stabilises the β-catenin pool, and it acts as a molecular sponge for miR-217-5p, relieving repression of HIF-1α. As both β-catenin and HIF-1α are master transcriptional regulators of glycolysis, their concerted up-regulation markedly increases the expression of glycolytic enzymes, boosts glucose consumption and lactate production, and accelerates tumour-cell proliferation. (Figure 2A) Thus, M2-TAM-derived exosomes enforce a glycolytic switch in GC cells via MALAT1-dependent co-activation of β-catenin and HIF-1α signalling, driving metabolic adaptation and malignant progression. 25 These findings identify the MALAT1/β-catenin/HIF-1α axis as a tractable therapeutic target for reversing glycolysis-associated chemoresistance and inhibiting tumour growth, offering significant translational potential.
3.2. M2-TAMs Derived Exosomes Drive Neovascularization, Thereby Accelerating Tumour Progression
PDAC ranks among the most aggressive human malignancies. 26 Yang et al documented a strong correlation between M2-TAMs and microvessel density in PDAC specimens. In vivo assays revealed that M2-TAM-derived exosomes markedly accelerate tumour growth and neovascularisation. Mechanistically, these vesicles shuttle miR-155-5p and miR-221-5p into endothelial cells, where the two miRNAs—via an E2F2-dependent transcriptional programme—promote angiogenesis. (Figure 2B) These findings delineate a critical regulatory axis underlying PDAC progression and offer a tractable therapeutic target for this lethal disease. 27
3.3. M2-TAMs Derived Exosomes Enhance Vascular Permeability, Thereby Facilitating Tumour Cell Metastasis
Late-stage metastasis accounts for the majority of deaths in patients with HCC,
28
dissecting the underlying mechanisms is therefore essential to extend survival. Lu et al showed that heterogeneous nuclear ribonucleoprotein A1 (hnRNPA1) recognises and binds pri-miR-23a-3p, accelerating exosome biogenesis and secretion. Exosomal miR-23a-3p directly targets phosphatase and tensin homologue (PTEN) and tight-junction protein 1 (TJP1); their down-regulation disrupts junctional integrity and hyper-activates pro-metastatic signalling. Moreover, exosome-educated HCC cells acquire a robust paracrine phenotype, secreting elevated levels of GM-CSF, VEGF, G-CSF, MCP-1/CCL2 and IL-4. These cytokines not only amplify neovascularisation and immunosuppression but also recruit additional monocytes/macrophages that polarise into M2-TAMs, thereby establishing a self-sustaining “M2-exosome → tumour progression → M2 recruitment” loop that drives metastatic progression.
29
Macrophage-derived exosomes internalised by tumour cells potentiate malignant progression. (A) TAM-released exosomes shuttle MALAT1 into gastric cancer cells, thereby accelerating proliferation and invasion. (B) In PDAC, TAM-derived exosomes deliver miR-221-3p and miR-155-5p to tumour cells, which repress E2F2 and promote intratumoural angiogenesis
4. Targeting Tumour-Associated Macrophages to Halt Cancer Progression
Based on the significant differences in cellular functional characteristics between M1-TAMs and M2-TAMs, next-generation TAM-directed therapeutics have been conceived to: Targeting macrophage receptors promotes the generation of M1-TAMs; block the recruitment of monocytes or pre-differentiated M2-TAMs; Promoting phagocytosis and antigen presentation of TAMs; Blocking signal transmission between tumor-associated macrophages and tumor cells. (Table 2)
4.1. Targeted Manipulation of Macrophage Receptors to Expedite the Transition From M2-TAMs to M1-TAMs
Compared with their M2 counterparts, M1-TAMs potently suppress neoplastic progression. Whether the two differentiation states—both originating from the M0 precursor—can be interconverted has therefore emerged as a pivotal determinant of therapeutic outcome. 30 Consequently, multiple strategies aimed at reverting the immunosuppressive M2-TAMs back to an M1-TAMs have been developed in recent years. 31
Toll-like receptors (TLRs) on macrophages sense tumour-derived molecular patterns and trigger cytokine secretion from TAMs. In the presence of interferon-γ, TLR agonists potently repolarise M2-TAMs toward an M1 phenotype in Lewis lung carcinoma(LLC) models. 32 The dual TLR7/8 agonist resiquimod (R848) has achieved clinically relevant M2-to-M1 reprogramming and elicited robust anti-tumour immunity. 33
Signalling via the M-CSF receptor (CSF-1R) and its ligands M-CSF and IL-34 is indispensable for M2-TAM specification. Pharmacological blockade of this axis therefore favours emergence of an M1-TAMs. The selective CSF-1R inhibitor BLZ945 has demonstrated potent anti-tumour activity in pre-clinical models, where it targets TAMs at multiple stages of mammary-to-brain metastasis and curtails tumour outgrowth. 34 Similarly, PLX3397 antagonises CSF-1R, abrogates p-ERK1/2 activation in bone-marrow-derived macrophages and attenuates M2-TAMs polarisation. Notably, PLX3397 also augments CD8+ T-cell infiltration within primary and metastatic osteosarcoma lesions, thereby amplifying anti-tumour immunity. 35
4.2. Inhibition of Monocyte/M2-TAM Recruitment Represents a Promising Therapeutic Strategy to Impede Cancer Progression
As monocytes constitute the predominant reservoir from which TAMs arise, intercepting their mobilisation or blocking the homing of pre-polarised M2-TAMs to the tumour micro-environment constitutes a rational therapeutic strategy. The two best-characterised axes are CCL2/CCR2 and CXCL12/CXCR4. Blockade of the CCL2–CCR2 axis sequesters monocytes within the bone marrow, thereby attenuating the generation of M2-TAMs at both primary and metastatic sites, while concomitantly augmenting CD8+ T-cell infiltration. 36 These convergent effects collectively suppress tumour growth and invasion. MK-0812, a selective CCR2 antagonist, potently disrupts CCL2–CCR2 engagement and abrogates downstream signaling activation, eliciting robust anti-tumour efficacy. 37 Additionally, activation of the HMGA2/STAT3/CCL2/CCR2 cascade promotes macrophage accumulation in colorectal cancer and accelerates tumour progression. 38 Consequently, CCR2 antagonists or anti-CCL2 monoclonal antibodies exhibit broad anticancer potential.
The CXCL12/CXCR4 axis governs the recruitment and subsequent differentiation of monocytes into tumour-associated macrophages, thereby fuelling tumour progression and metastasis.39,40 Small molecule antagonists (e.g., plerixafor), peptides/peptide mimetics (e.g., BKT140), monoclonal antibodies (e.g., PF-06747143 and ulocuplumab), and microRNAs have shown promise in reducing the tumor load, inducing apoptosis, and rendering malignant cells resistant to conventional chemotherapy. 41
4.3. Promoting Phagocytosis and Antigen Presentation of TAMs
Harnessing the potent phagocytic capacity of macrophages for tumour cell eradication represents a promising therapeutic strategy; however, the SIRPα-CD47 axis constrains macrophage-mediated recognition and engulfment. 42 SIRPα is expressed on macrophages, whereas its ligand CD47 is broadly expressed on both healthy and neoplastic cells. The SIRPα-CD47 axis transmits inhibitory signals that protect tumour cells from macrophage-mediated phagocytosis and clearance.43,44 Consequently, pharmacological disruption of this axis to restore phagocytic competence and suppress tumorigenesis represents an emerging therapeutic strategy in oncology. Current clinical-stage therapeutics targeting the SIRPα-CD47 axis comprise CD47-directed monoclonal antibodies (e.g., AK117, CC-90002, and Hu5F9-G4) and selective SIRPα antagonists, including SIRPα-IgG and SIRPα-Fc fusion proteins as well as SIRPα-targeting monoclonal antibodies. These agents potentiate macrophage phagocytosis through specific blockade of the SIRPα-CD47 interaction and have demonstrated preliminary anti-tumour efficacy across diverse malignancies. 45 Furthermore, advances in genetic engineering have propelled the development of innovative tumour-targeting modalities. Notably, genetically engineered cell membrane-coated magnetic nanoparticles (gCM-MNs) potently disrupt the SIRPα-CD47 axis. In preclinical melanoma and breast cancer models, gCM-MNs suppressed local tumour growth and abrogated distant metastasis, thereby substantially prolonging overall survival in murine subjects. 46
The SIRPα–CD47 axis represents an emerging therapeutic target that has demonstrated promising efficacy in haematological malignancies. In most solid tumours, selective SIRPα antagonists exhibit superior therapeutic activity compared with anti-CD47 monoclonal antibodies; nevertheless, their efficacy as monotherapy remains limited, and combinatorial regimens generally yield more favourable outcomes. Notably, given the ubiquitous expression of CD47 on haematopoietic cells, the administration of CD47-targeting agents necessitates vigilant monitoring for anaemia and thrombocytopenia. 45
4.4. Disrupting Intercellular Communication Between Neoplastic Cells and M2-TAMs Abrogates Tumour Progression
ATMs-Targeted Drugs
5. Discussion and Conclusion
This review focuses on exosome-mediated bidirectional signalling mechanisms between TAMs and malignant cells, elucidating the dynamic regulatory networks governing TAMs within the tumour microenvironment. (Table 1) Despite advancing mechanistic insights, pharmacological agents targeting these reciprocal interactions remain scarce, underscoring the imperative for innovative therapeutic development. Notably, these regulatory mechanisms predominantly highlight the pivotal role of transcriptional control in shaping macrophage identity and function. Emerging evidence indicates that post-transcriptional regulation exerts equally indispensable functions in macrophage reprogramming. 48 Post-transcriptional mechanisms, encompassing RNA-binding proteins (RBPs) and RNA modifications, exhibit accelerated kinetics and enhanced reversibility. 49 Dysregulation of RBPs or RNA modifications can alter macrophage polarisation and functional states, thereby influencing tumour progression. Specific RBPs promote M2-TAMs generation through interactions with circular RNAs (e.g., circFUT8-ENO1, circ-0003137-PTBP1, hsa-circ-0058495) and long non-coding RNAs (e.g., lncRNA-PACERR, lncRNA NR_109), fostering an immunosuppressive microenvironment.50-54 RNA modifications, including N6-methyladenosine (m6A) and N4-acetylcytidine (ac4C), modulate macrophage polarisation and activation to varying extents. In nasopharyngeal carcinoma, m6A-modified WNT2 activates FZD2/β-catenin signalling in macrophages to induce M2-TAMs, promoting tumour growth and metastasis. 55 In atherosclerosis, NAT10 catalyses ac4C modification of TLR9 mRNA to regulate M2-TAMs polarisation, accelerating disease progression. 56 Given the fundamental importance of post-transcriptional regulation in macrophage functional control, targeting RBPs or RNA modifications represents a promising avenue for immunotherapeutic investigation.
Furthermore, the traditional M1/M2 dichotomy represents an oversimplified framework. The polarization of TAMs is governed by a complex interplay of intracellular signaling pathways and extracellular environmental factors within the TME, constituting a dynamic and highly plastic process. 57 The advent of single-cell sequencing and multi-omics technologies has substantially advanced our understanding of the molecular heterogeneity of TAMs. Ma et al further stratified TAMs into seven distinct subpopulations: interferon-stimulated (IFN-TAMs), immune-regulatory (Reg-TAMs), inflammatory cytokine-enriched (Inflam-TAMs), lipid-associated (LA-TAMs), pro-angiogenic (Angio-TAMs), tissue-resident macrophage-like (RTM-TAMs), and proliferative TAMs (Prolif-TAMs), thereby underscoring the limitations of the binary classification and the dynamic diversity of TAMs. 58 Moreover, TAM phenotype and function are modulated by tumor developmental stage, intratumoral spatial location, and tumor type. In hepatocellular carcinoma, TAMs exhibit non-uniform spatial distribution, predominantly presenting an M2-TAMs in invasive regions at the tumor periphery. 59 Conversely, in gastric cancer, M1-TAMs predominate in the tumor core, whereas M2-TAMs are primarily localized at the tumor margin 60 . Understanding TAM heterogeneity holds significant implications for cancer therapeutics; given that TAMs demonstrate context-dependent phenotypic and functional plasticity across temporal and spatial dimensions, treatment strategies must evolve to accommodate the specific phenotypic spectra and functional characteristics observed in distinct tumor contexts.
Therapeutic targeting of TAMs, propelled by emerging technologies, is forging innovative and clinically promising avenues in oncology. Nevertheless, substantial challenges persist. First, monotherapies directed against TAMs demonstrate limited efficacy, while treatment-related adverse events remain a significant concern; consequently, the development of novel combinatorial strategies is imperative.45,61 Second, insufficient targeting specificity constitutes a major limitation of current agents; the advancement of highly specific TAM-directed therapeutics—exemplified by nanoparticle-based drug delivery systems—represents a viable solution.62,63 Third, the temporal and spatial heterogeneity of TAMs poses formidable obstacles to drug development; therefore, refined subpopulation stratification and the identification of more distinctive therapeutic targets warrant intensive investigation. 64 Finally, the translational gap between preclinical findings and clinical efficacy warrants explicit acknowledgment. Current preclinical platforms—by virtue of their reductionist design—exhibit homogeneous cellular compositions and lack the intricate stromal–immune–tumour interactions that govern therapeutic response in vivo. While these models yield indispensable mechanistic insights, human malignancies present substantially greater cellular complexity and dynamic intercellular crosstalk that remain imperfectly recapitulated. Consequently, the progression from preclinical proof-of-concept to clinical validation necessitates iterative optimisation and rigorous phase-specific testing.
Footnotes
Ethical Considerations
This article is a review article. The content does not involve any clinical human samples or animal experiments, and therefore does not raise ethical concerns.
Author Contributions
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This review is supported by a key project of Nanjing Medical Technology Development Fund (No.ZKX19046)
Declaration of Conflicting Interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Data Availability Statement
No data was used for the research described in the article
Statement on AI Usage in This Article
Regarding the use of AI in this manuscript, we hereby make the following statement: We solemnly pledge that no AI tools were employed to generate any content in this manuscript, and none of its contents were completed with the assistance of AI tools. After completing the manuscript, we used the DeepSeek tool to verify the appropriateness of word and syntactic usage in individual sentences; beyond this, no other AI tools were utilized for further editing or processing of the manuscript.