TAMs exhibit a high capacity for glucose uptake⁶³. This glucose is not merely used for energy production; glucose metabolites also regulate proteins through post-translational modifications and directly influence TAM phenotype and function by altering metabolic enzyme activity. Furthermore, tumor cells and TAMs engage in a vicious cycle of glucose metabolism that drives tumor progression. In response to stimulation, the activation of pathogen-associated patterns, such as Toll-like receptor (TLR4) or lipopolysaccharide (LPS) stimulation, activate the PI3K/AKT pathway, which leads to upregulation of glycolysis and lactate production, resulting in a pro-inflammatory TAM phenotype^17,64,65. IL-4 or IL-13 shifts the metabolic program of macrophages toward OXPHOS and FAO in the TME. This transition depends on AMPK activation, which promotes FAO and enhances mitochondrial function. Concurrently, mechanistic target of rapamycin complex 2 (mTORC2) and the Akt–mTORC1 signaling axis are also engaged, although the roles are more selective. For example, IL-4 signaling upregulates ATP-citrate lyase activity through the Akt–mTORC1 pathway, thereby increasing acetyl-CoA levels. This increase in acetyl-CoA facilitates histone acetylation and enhances expression at specific M2 gene loci⁶⁶. By contrast, mTORC1 activation usually upregulates HIF-1α. As a transcription factor, HIF-α then promotes the expression of glycolytic enzymes, such as lactate dehydrogenasae A (LDHA) and hexokinase 2 (HK2). This enables TAMs to undergo glycolysis, even in the presence of oxygen⁶⁷. Intreastingly, TAMs have been reported to promote glioblastoma cell survival by releasing extracellular vesicles that contain LDHA⁶⁸. Such glycolysis-dependent metabolic reprogramming in macrophages is essential for maintaining metabolic homeostasis. Moreover, glycolytic enzymes and metabolites support the functions of TAMs through intrinsic and moonlighting mechanisms. Hypoxic TAMs accumulate HIF-1α in lung carcinomas, which upregulates HIF-1α target genes encoding glycolytic enzymes (Slc2a1 and ldha)⁶⁹. These glycolytic enzymes, e.g., pyruvate kinase M2 (PKM2) and LDHA, are highly expressed in TAMs, sustaining the TAM functional phenotype transition through distinct mechanisms^70–72. PKM2 is the rate-limiting enzyme in glycolysis that converts phosphoenolpyruvate (PEP) and adenosine diphosphate to pyruvate and adenosine triphosphate. The direct binding of PKM2 and HIF-1α to hypoxia response element sites in the PD-L1 promoter promotes PD-L1 expression under LPS stimulation in macrophages⁷³. Myeloid-specific deletion of LDHA accumulated M1 phenotype macrophages in a K-Ras murine model of lung carcinomas, which was characterized by CD86^high and CCL2^high, and suppressed tumor growth⁷⁴.

In addition, the TME soluble factors, such as catabolites, cytokines, and exosomes, have critical roles in modulating TAM glycolysis and coordinating TAM functions. Macrophages located in the peritumoral region exhibit increased glycolysis in hepatocellular carcinoma (HCC) compared to non-tumor liver tissue or circulating monocytes. Mechanistically, the critical regulator of glycolytic flux rate, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3), is the major driver of this glycolytic surge, which leads to p65 nuclear translocation and activation of the NF-κB pathway in monocytes. Secreted TNF-α/IL-10 is subsequently elevated and interacts in a loop with receptors on these peritumoral monocytes to increase the expression of the immune checkpoint molecule, PD-L1⁷⁵. Tumor-drived exosomes (TDEs) trigger glycolytic metabolic reprogramming in TAMs and help to maintain the immunosuppressive phenotype. TDEs enhance glucose uptake through NF-κB and the TLR2-dependent pathway, and inhibit mitochondrial OXPHOS to convert pyruvate-to-lactate through nitric oxide synthase 2 (NOS2). The end product, lactate, then feeds back to NF-κB, further increasing PD-L1 expression¹⁷. In addition to the TME, chronic exposure to environmental factors has been shown to maintain the functional phenotype of TAMs through metabolic reprogramming. It has been shown that smoking tobacco contributes to lung cancer progression through multiple mechanisms, including metabolic reprogramming (e.g., induction of GLUT1) and immune suppression. One potential pathway involves upregulation of factor insulin-like growth factor 2 (IGF2) in the TME, which activates receptors, such as insulin receptors (IRs) on TAMs, leading to increased PD-L1 expression and subsequent inhibition of anti-tumor immunity⁷⁶. Obesity is the common factor of tumor progression. The cytokines and molecules linked to obesity (e.g., leptin, insulin, palmitate, IFN-γ, and TNF) upregulate programmed death-1 (PD-1) expression on TAMs in an mTORC1- and glycolysis-dependent manner. PD-1 acts as a self-inhibitory receptor that suppresses the glycolysis, phagocytosis, and antigen presenting ability of TAMs⁷⁷.

The end product of glycolysis, lactate, functions as a signaling molecule and facilitates TAM glucose metabolism to support TAM function through several pathways. Tumor-derived lactic acid, as a soluble factor in the TME, stabilizes HIF-1α and induces Vegf/Arg1 expression in TAMs through integrated metabolic, receptor-mediated, and epigenetic mechanisms. This axis represents a therapeutic vulnerability in cancer⁷⁸. While glucose and glutamine also feed into the TCA cycle, lactate becomes dominant under conditions, such as high extracellular lactate (e.g., the TME). With LPS stimulation under these conditions, lactate-derived acetyl-CoA upregulates acetylation of the histone, H3K27, and this epigenetic modification opens chromatin regions, enabling the transcription of immunosuppressive genes (e.g., Nuclear Receptor subfamily 4 group A member 1). This mechanism may be hijacked in tumors to evade immunity⁷⁹. Monocyte-derived macrophages (MDMs) outnumber resident microglia (MGs) in advanced glioblastoma multiforme and become the primary immunosuppressive macrophage population. De Leo et al.⁸⁰ reported that tumor-derived factors (e.g., cytokines and metabolites) activate the protein kinase r-like endoplasmic reticulum kinase (PERK)-activating transcription factor 4 (ATF4) axis in myeloid-derived macrophages (MDMs), inducing glucose transporter GLUT1 expression and hyperglycolysis. Lactate modifies histones via lactylation [lysine lactylation (Kla)], particularly on the promoters of immunosuppressive genes (e.g., IL-10), thus inhibiting T cell function and furthering tumor growth. Li et al.⁸¹ uncovered how lactate in the TME of colorectal cancer (CRC) epigenetically educates TAMs. Specifically, tumor-derived lactate is taken up by TAMs, which fuels H3K18 lactylation and thuereby suppresses retinoic acid receptor-gamma (RARγ) expression in TAMs, activating NF-κB. This process results in IL-6 production, which activates janus kinase/signal transducer and activator of transcription 3 signalling in CRC cells to promote colorectal tumor growth, progression, and metastasis. This finding identifies a critical metabolic-epigenetic-immune axis driving CRC.

TAMs, specifically pro-tumor M2-like TAMs, exhibit significantly increased glucose uptake to fuel the hexosamine biosynthetic pathway (HBP). The HBP produces UDP-GlcNAc, the donor substrate for a post-translational modification [O-linked β-N-acetylglucosaminylation (O-GlcNAcylation)], which is catalyzed by the enzyme, O-GlcNAc transferase (OGT). O-GlcNAcylation targets cathepsin B (CTSB). Stabilized CTSB is then secreted by TAMs into the surrounding TME to facilitate tumor cell detachment and migration⁸². Glycolysis co-opts with the TCA cycle in the macrophage phenotype transition.

In addition to enhanced glycolytic flux, studies have revealed a metabolic “break” within the TCA cycle. In pro-inflammatory M1-like states, the TCA cycle is disrupted by isocitrate dehydrogenase (IDH). Under normal conditions, isocitrate is converted to α-ketoglutarate (α-KG) in the TCA cycle, after which α-KG undergoes OXPHOS. This break, also known as “TCA cycle rewiring”, involves the diversion of citrate away from oxidative metabolism. Instead, aconitate decarboxylase 1 (Immune-responsive gene 1 (IRG1), encoding aconitate decarboxylase 1) converts cis-aconitate to itaconate, a potent inhibitor of succinate dehydrogenase (SDH), with immunomodulatory properties^26,83. This conversion causes succinate buildup, stabilizing HIF-1α and inducing IL-1β secretion⁸⁴. As shown, significant disruption in TCA cycle function occurs in TAMs, driven by the unique metabolic demands of the TME. These disruptions fundamentally alter TAM function, promoting immunosuppression and tumor progression.

It has been shown that itaconate is the most highly upregulated metabolite in peritoneal tissue-resident macrophages (pResMΦs) from peritoneal tumors (B16 melanoma or ID8 ovarian carcinoma). Itaconate drives FAO to promote OXPHOS, inducing mitochondrial reactive oxygen species (mtROS) production. mtROS produced by pResMΦs diffuse into neighbouring tumor cells, activating MAPK signalling and ultimately favor tumor cell proliferation and survival⁸⁵. Chen et al.⁸⁶ demonstrated that tumor cells activate NF-κB signalling to elevate Irg1 expression in TAMs. Itaconic acid (ITA) produced by ACOD1 suppresses inflammatory gene expression and CD8⁺ T cell infiltration within tumor sites by inhibiting ten-eleven translocation DNA dioxygenases. These findings unequivocally identify itaconate as a pro-tumor metabolite. The role of itaconate in driving TAM-mediated immunosuppression makes itaconate a potential therapeutic target for cancer.

In addition to itaconate, cancer cells (particularly lung cancer cells) release high levels of succinate into the TME resulting from reduced succinate dehydrogenase complex iron sulfur subunit B (SDHB) expression or SDH mutation. This process causes succinate buildup, stabilizing HIF-1α and inducing IL-1β secretion⁸⁴. Expression of the succinate receptor 1 (SUCNR1) is also increased in tumor tissue. Extracellular succinate binds to its receptor (SUCNR1) on TAMs to activate the PI3K-HIF1α axis, which stabilizes HIF, thus promoting VEGF secretion and inhibiting T cell function⁸⁷. Zhu et al.⁵ reported that the ketogenesis gene, 3-oxoacid CoA-transferase 1, is highly expressed in TAMs and promotes a pro-tumor phenotype in TAMs by suppressing CD8⁺ T cell exhaustion. Mechanistically, OXCT1-mediated ketolysis generates succinate as a byproduct in TAMs. Succinate elevates H3K4me3 (histone trimethylation) levels at the promoter of Arg1. This change upregulates Arg1 expression, polarizing TAMs towards a pro-tumor phenotype. Furthermore, Lu et al.³⁷ revealed that tumor-derived microparticles loaded with succinate (SMPs) remodel polarization of M1-like macrophages by enhancing glycolysis and attenuating the TCA cycle in a protein succinylation–dependent manner. Succinate accumulation from SMPs induces mitochondrial IDH2 succination, which stabilizes HIF-1α. In contrast, succinate induces histone H3K122 succinylation at the LDHA locus in TAMs, which enhances glycolysis and pro-inflammatory polarization, ultimately potentiating anti-tumor immunity.

The concurrent upregulation of glycolysis and reprogramming of the TCA cycle are both essential metabolic adaptations for inflammatory macrophage activation. Upon stimulation with LPS or within the TME, TCA cycle reprogramming leads to the accumulation of specific intermediates, such as itaconate and succinate. These metabolites function as immunomodulatory signalling molecules and participate in inflammatory responses through distinct mechanisms. Concurrently, disruption of the TCA cycle at specific points creates a dependency on upregulated glycolysis. This glycolytic shift provides the necessary ATP for the cellular energetics and biosynthetic precursors required to sustain macrophage effector functions during inflammation. The glucose metabolism-related signaling pathways in TAMs are summarized in Figure 2.

Lipid metabolism is central to regulating TAM biology. Lipid metabolism governs the energy supply and the phenotypic state of TAMs through fatty acid metabolism, while specific lipid derivatives function as key signaling molecules. Notably, recent research has revealed that lipid metabolism in TAMs can also act as a “metabolic supplier” for tumor cells⁸⁸. Lipid perturbations impact the immune microenvironment, including TAMs, and have been verified in various cancer types, such as breast cancer, lung cancer, and clear cell renal cell carcinoma (ccRCC). As a prototypical tumor characterized by metabolic reprogramming, the metabolic heterogeneity in the ccRCC TME has recently been well-documented. In the current study macrophages were reported to undergo varying polarization states across distinct metabolic states. For example, FATTY⁺ macrophages (demonstrating increased fatty acid synthesis metabolic pathway activity) and amino acid⁺ macrophages (exhibiting increased amino acid and glutathione metabolic pathway activity) emerge within pro-inflammatory macrophage subtypes⁸⁹.

Lipid metabolism in macrophages is essential for phenotypic reprogramming and considerable progress has been made in elucidating the underlying mechanisms. For example, studies in mouse models have shown that a high-fat diet increases the abundance of lipid-loaded TAMs in the TME, which release CCL6 (the human functional homologue is CCL23) to promote cancer cell migration in prostate adenocarcinoma (PCa). This process is driven by increased expression of the receptor, macrophage receptor with collagenous structure [macrophage receptor with collagenous structure (MACRO) C-type lectin domain family 5 member A (CLEC5A)], on macrophages, which is triggered by tumor cell-derived IL-1β. MACRO functions as a scavenger receptor that governs macrophage lipid uptake. In the current study blocking MACRO with an antibody suppressed tumor growth and improved the efficacy of chemotherapy in PCa. This approach may provide a promising anticancer strategy for patients with abnormal lipid metabolism⁹⁰. Peroxisome proliferator-activated receptor-gamma (PPARγ) is a nuclear hormone receptor that senses cellular fatty acid levels and coordinates lipid transport by regulating fabp5 gene expression. PPARγ has been reported to be involved in the function of fatty acid-binding protein 5 (FABP5⁺) lipid-loaded TAMs. Specifically, tumor cells secrete long-chain unsaturated fatty acids, including docosahexaenoic acid (DHA), which are taken up by TAMs. FABP5 transports these fatty acids to the nucleus inside TAMs, leading to activation of PPARγ. This activation drives the upregulation of immunosuppressive genes, like PD-L1, PD-L2, and Gal-1. By suppressing effector T cell proliferation and activation, this pathway ultimately facilitates HCC progression⁹¹. In addition, pyruvate kinase M2 (PKM2) drives the Warburg effect in macrophages and interacts with HIF-1α to upregulate the expression of glycolytic enzymes, such as GLUT1 and LDHA⁹². Tumoral exosomal PKM2 can polarize TAMs to an anti-inflammatory phenotype to enhance the proliferation, migration, and invasion of gastric cancer (GC) cells. Mechanistically, exosomal PKM2 increases the accumulation of sterol regulatory element-binding protein 1 (SREBP1) in the nucleus to upregulate fatty acid synthesis-related enzymes [e.g., fatty acid synthase (FASN), acetyl-CoA carboxylase alpha (ACACA), and ATP citrate lyase (ACLY)]⁹³. In addition to lipid synthesis, phenotype rewiring in TAMs can also be achieved through lipid catabolism and FAO. Tian et al.⁹⁴ identified glucose-regulated protein 78 (GRP78), which is secreted by colon cancer cells, as a regulator of TAM polarization. GRP78 enters macrophages and localizes to lipid droplets, enhancing the stability of adipose triglyceride lipase (ATGL). This process polarizes macrophages towards an anti-inflammatory phenotype mediated by interactions among AA, DHA, and PPARγ⁹⁴. As the tumor progresses, the predominant subtype of TAMs becomes M2-like, which is accompanied by improved efficacy of FAO. Blocking FAO with perhexiline (PerHx) can hinder anti-inflammatory phenotype transition of TAMs⁹⁵. Several mechanisms have been proposed to be involved in this process. FAO provides the energy and metabolic intermediates needed for TAMs to suppress anti-tumor T cell activity (e.g., through arginase production, PD-L1 expression, and the secretion of cytokines, such as IL-10), thus inhibiting the immune function of T cell. In contrast, JAK1 functions as an activator of pro-inflammatory signals, whereas SHP-1 acts as a negative regulator to prevent excessive activation. The opposing actions converge within the JAK–STAT pathway to fine-tune macrophage activity². In this context, elevated FAO contributes to the promotion of mitochondrial OXPHOS, production of reactive oxygen species, phosphorylation of JAK1, and dephosphorylation of src homology region 2 domain-containing phosphatase 1, leading to STAT6 activation and transcription of genes (especially C-C Motif Chemokine Ligand 2 and Arg1) that regulate TAM generation and function²¹. Furthermore, several studies have revealed that the scavenger receptor, CD36, has a vital role in the macrophage transition to a more pro-tumoral state. For example, lipid-loaded vesicles are absorbed by macrophages via CD36. Genetic ablation of CD36 in these macrophages using LyzM-Cre mice resulted in decreased expression of the M2-like marker, CD206, and a concomitant increase in the M1-like marker, CD80. This shift was associated with the activation of CD8⁺ T cells and the attenuation of liver metastasis^40,96. Interestingly, CD36 does not affect the CCL2 level, a chemoattractant produced by cancer cells for MDMs but influences how these macrophages react to CCL2. Mechanistically, the myeloid cells highly express p110γ. CD36 alters the sphingolipid composition of lipid rafts, promoting activation of the lipid raft-dependent PI3K catalytic subunit, p110γ. This effect leads to a heightened response to the CCL2/C-C chemokine receptor type 2 chemotactic signal from tumor cells. The use of p110γ inhibitors in mouse models of liver metastasis can inhibit monocyte infiltration, restore anti-tumor immunity, and eliminate the effects of CD36 deficiency on the TME⁹⁶. Nian et al.⁹⁷ reported that cancer stem cells (CSCs) establish an IL-34-orchestrated niche to promote tumorigenesis in p53-inactivated liver cancer. The authors discovered that p53 inactivation in tumors induces IL-34 secretion, after which IL-34 upregulates CD36 in TAMs, enhancing lipid uptake and switching metabolism to FAO. This process drives TAMs towards a foam-like, M2-like phenotype. These reprogrammed TAMs are necessary and sufficient to suppress CD8⁺ T cells, enabling the tumor to escape immune surveillance. Targeting the IL-34-CD36 axis disrupts this pathway, overcomes immune escape, and inhibits tumor growth, particularly in p53-deficient cancers. This finding identified the IL-34-CD36 axis as a crucial mechanism and therapeutic vulnerability for immune-evasive cancers with p53 mutations⁹⁷.

The alternatively-activated state of TAMs is usually induced by IL-4. Notably, membrane cholesterol efflux in TAMs is increased in a mouse model of metastatic ovarian cancer, thus promoting IL-4-induced pro-tumor TAM reprogramming in a STAT6/PI3K-dependent manner⁹⁸. When cholesterol-using agents, such as apolipoprotein A1 (ApoA1) were depleted or macrophages were treated with conditioned medium from ID8 cells, the lipid density decreased. This effect was accompanied by a reduction in IFN-γ-induced Nos2 expression and an increase in the expression of the M2-like marker, Arg1. However, Wang et al.²⁵ identified a positive correlation between cholesterol levels and “do not eat me” signaling in MDMs. The work in GBM revealed that increasing cholesterol efflux via the transporter, APOA1, can reinstate macrophage phagocytosis and restore anti-tumor immunity mediated by macrophages and T cells. The dual role of cholesterol efflux highlights the importance of tumor-specific metabolic targeting in immunotherapy. Several lipids have been identified as key regulators of TAM function in the TME. For example, the TME often shows elevated 25-hydroxycholesterol (25-HC) levels. This oxysterol, often produced during inflammation or infection, actively drives TAMs towards an immunosuppressive phenotype and contributes to cold tumors^99,100. The underlying mechanism has been elucidated recently. A cholesterol-25-hydroxylase (CH25H)-high subset of macrophages is enriched among immunosuppressive TAMs. Lysosome-accumulated 25-HC competes with cholesterol for binding to G protein-coupled receptor 155, thereby inhibiting the kinase mTORC1 and activating the energy sensor, AMPK. Activated AMPK orchestrates a metabolic shift in TAMs, specifically promoting OXPHOS and reducing glycolysis. This metabolic switch is essential for 25-HC to educate macrophages in a STAT6-dependent manner, providing a mechanistic link between cholesterol metabolism, cellular energy sensing, and immune evasion, particularly in colorectal cancer¹⁰¹. In addition to cholesterol, AA and its metabolic product, PGE₂, have been well-documented as immunomodulators that educate macrophages in the TME. Hammoud et al. specifically revealed a non-metabolic function of AA within the ovarian cancer immune microenvironment. The study showed that AA acts as an adverse mediator of JAK-STAT signalling in immunosuppressive TAMs by displacing key signalling proteins, including IFNAR1, STAT1, and JAK1, out of cholesterol-enriched lipid raft microdomains, thus promoting an immunosuppressive niche that facilitates tumor progression¹⁰². Another study highlighted the crucial role of AA in SLC3A2-mediated macrophage polarization in lung adenocarcinoma¹⁰³. PGE₂ activates downstream signaling pathways, including the PI3K–AKT and MAPK signaling pathways, via its receptors, such as E-type prostanoid receptor (EP) 2/4, thereby modulating macrophage function. Notably, PGE₂ can profoundly alter mitochondrial activity in macrophages. For example, in IL-4-induced macrophages, PGE₂ downregulates key genes involved in the malate–aspartate shuttle. This reduction leads to a decrease in mitochondrial membrane potential, which in turn triggers a cascade of changes in nuclear gene expression and ultimately regulates cellular function¹⁰⁴. As an important immunosuppressive mediator in the TME, PGE₂-EP/EP4 signalling impairs both innate and adaptive immunity in TME. For example, a distinct subset of TAMs in pancreatic ductal adenocarcinoma (PDAC) tumor samples expressing high levels of IL-1β is correlated with poor prognosis. Activated PDAC cells produce more PGE₂ and TNF-α, reinforcing IL-1β⁺ TAM differentiation and IL-1β from TAMs further activate PDAC cells to produce more PGE₂ (via prostaglandin-endoperoxide synthase 2/Cyclooxygenase-2) and TNF-α, resulting in a self-reinforcing cycle¹⁰⁵. In addition, PGE₂ is well-established as an inhibitor of OXPHOS in macrophages, particularly M1-like macrophages. Recently, PGE₂ treatment was reported to suppress OXPHOS and ribosomal protein (RP) biogenesis via suppression of peroxisome proliferator-activated receptor gamma coactivator 1αα and cell myelocytomatosis oncogene signaling in macrophages. EP4 receptor expression is elevated in activated macrophages, and when present, PGE₂ acts predominantly through this receptor to suppress the pro-inflammatory activity of M1-like macrophages in the TME¹⁰⁶. Cilenti et al. unravelled a PGE₂-mediated pathway that fine-tunes inflammatory responses in macrophages, which present a transcriptional circuit for IFN-I induction relevant to cancer. PGE₂ interferes with the LPS-mediated activation of extracellular signal-regulated kinase 5, a kinase that acts as a transcriptional partner of myocyte-specific enhancer factor 2A (MEF2A). This disruption prevents ERK5 from phosphorylating/activating MEF2A, thereby dampening IFN-I responses. Consequently, reduced IFN-β leads to diminished expression of a series of pro-inflammatory genes¹⁰⁷. This evidence supports an immunosuppressive function of PGE₂, especially in response to the pro-inflammatory stimuli in the TME. Similarly, fatty acid and triglyceride metabolism appear to be mechanistically crucial in the functional transition of TAMs¹⁰⁸. TAMs influence tumor progression by modulating other cellular components within the TME through the secretion of PGE₂. As previously reported, M2-like macrophages promote the differentiation of immunosuppressive cells, such as regulatory T cells (Tregs) and regulatory B cells (Bregs)¹⁰⁹. CD169⁺ resident TAMs from lymph nodes, which have been associated with a favorable prognosis in breast cancer patients, significantly inhibited the proliferation of T cells, resting B cells, and NK cells in in vitro co-culture experiments. Conversely, CD169⁺ resident TAMs from lymph nodes markedly enhanced the antibody-producing capacity of activated B cells. These effects were linked to soluble mediators secreted by TAMs, including PGE₂ and IL-10¹¹⁰. Sphingolipid metabolism has recently emerged as a key process that links obesity, chronic inflammation, and breast cancer progression by polarizing macrophages toward pro-inflammatory phenotypes and recruiting TAMs to establish a metastasis-favoring TME¹¹¹. LPS/TLR4 activation induces de novo sphingolipid synthesis and this metabolic shift is distinct between pro- and anti-inflammatory macrophages. Specifically, sphinganine physically interacts with the TLR4 adaptor proteins, myeloid differentiation primary response 88 and TIR domain containing adaptor protein, facilitating recruitment to the cell membrane after TLR4 activation by LPS. This interaction is essential for initiating downstream NF-κB signalling and pro-inflammatory cytokine production. Inhibition of de novo sphingolipid synthesis, and thus sphinganine production in macrophages has been shown to impair the ability to recognize tumor-derived factors via pattern recognition receptors, such as TLRs. This impairment hinders the potential to switch towards a more anti-tumor state in melanoma mouse models¹¹². As the central intermediate in sphingolipid metabolism, ceramide in TAMs is essential for lipid droplet induction and fatty acid catabolism. Sun et al. reported that the expression of neutral ceramidase (nCDase) is significantly upregulated in TAMs within breast cancer. nCDase hydrolyzes ceramide into sphingosine, which is then phosphorylated to sphingosine-1-phosphate (S1P). This enzymatic activity shifts the sphingolipid balance in TAMs. By conditionally knocking out nCDase in macrophages or supplementing the culture medium with ceramide, the nCDase-S1P axis was shown to promote the differentiation and accumulation of a specific immunosuppressive triggering receptor expressed on the myeloid cells-2 (TREM2)⁺ TAM subset. These TREM2⁺ macrophages, known for pro-tumor functions, drive breast cancer progression by inducing CD8⁺ T cell exhaustion¹¹³. Within the TME, cancer-associated fibroblasts (CAFs) have an initiating role by “educating” and activating TAMs. This crosstalk is mediated through CAF-secreted C-X-C motif chemokine ligand 12 signaling via the C-X-C motif chemokine receptor 4 on TAMs. Activation by CAFs induces marked lipid metabolism reprogramming in TREM2⁺ lipid-associated macrophages (LAMs). The consequent immunosuppressive milieu further reinforces CAF activation, establishing a feed-forward loop that accelerates tumor progression. Supporting this, animal experiments in the study by Cao et al. demonstrate that the concurrent presence of CAFs and TREM2⁺ LAMs exerts robust tumor-promoting effect, driving both primary gastic tumor growth and lung metastasis¹¹⁴. Figure 3 summarizes the lipid metabolism-associated signaling pathways in TAMs.

TAMs utilize amino acid metabolism to directly support an immunosuppressive state within the TME by consuming key amino acids, producing immunosuppressive metabolites, and interacting with tumor cell metabolism¹⁵. For example, glutamine is suggested to drive the M2-like phenotype in macrophages¹¹⁵. Glutamine-derived metabolites, e.g., α-ketoglutarate, succinate, and aspartate, are identified as critical regulators of macrophage functions through metabolic and epigenetic reprogramming¹¹⁶. Orillion et al. reported that amino acid deprivation suppresses TAM infiltration and tumor growth, while promoting sensitivity to immunotherapies. In the current study dietary amino acid restriction was shown to drive macrophage repolarization to a tumor-suppressive state, characterized by altered cytokine production (e.g., increased IL-12/TNF-α and decreased IL-10) and enhanced phagocytosis via the ROS-mTOR cascade¹¹⁷. In addition, CAFs synthesize and release glutamine, which is subsequently taken up by TAMs. This uptake promotes the polarization of TAMs towards a pro-tumor phenotype, thereby contributing to an immunosuppressive TME and supporting tumor growth. Importantly, CRISPR/Cas9-mediated knockdown of glutamine synthesis (GS) in CAFs significantly impaired the capacity to induce macrophage polarization, demonstrating that the glutamine-producing function of CAFs is critical for driving this process¹¹⁸.

Enhanced glutamine metabolism has been implicated as an important switch that governs TAM polarization within the TME. GS expression and activity are induced by anti-inflammatory stimuli within the TME, especially under starvation, and GS abrogation makes macrophages reliant on extracellular glutamine uptake, which activates the mTORC1 signalling pathway, thus resulting in macrophage reprogramming towards an M1-like phenotype¹¹⁹. Macrophages in myeloid-rich prostate and bladder cancers (known as immunologically ‘cold’ and myeloid-rich tumors) treated with the glutamine antagonist, JHU083, or reprogrammed macrophages exhibited increased glycolysis, disrupted TCA cycle, and disrupted purine metabolism. Moreover, in addition to pro-tumor polarization, this metabolic rewiring might promote phagocytosis via myosin 1e. The myosin family protein, Myo1E, which is encoded by the gene, Myo1e, is essential for adhesion turnover during phagocytosis in macrophages. Myo1E facilitates membrane-cytoskeleton crosstalk to enable phagocytic cup closure¹²⁰. This effect of glutamine metabolism on phenotype skewing exists not only in TAMs but also in other suppressive myeloid cells, such as myeloid-derived suppressor cells (MDSCs)¹²¹. Recently, Liu et al. demonstrated that CD40 signaling promotes pro-inflammatory macrophage polarization by enhancing FAO and glutamine metabolism in melanoma. First, bone marrow-derived macrophages were treated with the CD40-activating monoclonal antibody, FGK45. This treatment increased the oxygen consumption rate (OCR) and triggered glycolysis in a glucose-independent manner. When OXPHOS was inhibited using rotenone or oligomycin, FGK45-induced M1-like polarization was blocked. Second, using Carnitine palmitoyltransferase 1-deficient mice to inhibit FAO or by knocking out glutaminase 1 (GLS1)/LDHA to disrupt glutamine metabolism, CD40-triggered macrophage polarization was similarly suppressed. Mechanistically, fatty acids are broken down into acetyl-CoA during FAO. Using ¹³C-labeled palmitate tracer analysis, FGK45 treatment significantly increased the levels of citrate and other TCA cycle metabolites in macrophages. Citrate serves as an acetyl-CoA donor for protein acetylation and epigenetic regulation. Moreover, glutamine acts as a substrate for glycolysis, which in turn fine-tunes the NAD⁺:NADH ratio to maintain NAD⁺ levels and support FAO. It is now clear that enhanced glycolysis stabilizes HIF-1α, with breakpoints at IDH and SDH in the TCA cycle during LPS-induced pro-inflammatory polarization¹²². Intriguingly, CD40-mediated M1 polarization occurs in an OXPHOS-dependent manner, and notably, the TCA cycle remains intact during this process. CD40-mediated M1 polarization reveals a novel and previously unreported mechanism of metabolic reprogramming underlying macrophage polarization triggered by CD40¹²³. Extracellular vesicles (EVs) from tumor cells, particularly microvesicles containing GLS1, is another factor to facilitate glutamine metabolism, then drive the M2-like polarization of TAMs. Hu et al. reported that HER2-positive gastric cancer cells (trastuzumab-resistant) exhibit increased CDC42 expression. Treatment with the CDC42 inhibitor, ZCL278, significantly reduced microvesicle marker protein expression. Mechanistically, CDC42 activated NF-κB p65 to regulate GLS1, promoting GLS1 microvesicle secretion through IQ motif containing GTPase activating protein 1. In co-culture experiments with macrophages, the addition of the GLS1 inhibitors, BPTES or CB839, significantly decreased GLS1 expression in macrophages. This treatment also upregulated M1 macrophage markers and reduced M2 macrophage markers¹²⁴. In addition, glutamine-derived aspartate contributed to the pro-inflammatory phenotype reprogramming of TAMs in HCC. As reported, aspartate upregulates spermidine through the polyamine synthesis pathway. Spermidine facilitates Eukaryotic translation initiation factor 5A hypusination, which contributes to the efficient translation of mRNAs encoding proteins with polyproline motifs, including HIF-1α¹²⁵. Mills et al. revealed that one of the breaks in the Krebs cycle involved in pro-inflammatory macrophage polarization and the increased IL-1β level caused by SDH also requires HIF mRNA translation and stabilization¹²⁶.

Immunosuppressive TAMs are characterized by tryptophan metabolism skewing, which is evidenced by enhanced indoleamine 2,3-dioxygenase 1 (IDO1) and/or aryl hydrocarbon receptor (AhR) activity¹²⁷. In a study integrating 178,651 human mononuclear phagocytes from 13 tissues, IL4I1⁺ MDMs were shown to have a higher level of the IDO1 gene in tumors than adjacent tissues and exhibit immunosuppressive characteristics¹²⁸. The monocytes gathered at the tumor periphery in a T cell-dependent manner and were matured by IFN-γ (produced via IFN-γ signalling and the CD40/CD40L interaction). The matured TAMs then recruit Tregs to tumors through degradation of tryptophan. In addition to TAMs, studies of other tumoral intrinsic components in classic Hodgkin lymphoma (cHL) have demonstrated that mononuclear phagocytes (e.g., cDCs and monocytes) also express immunoregulatory factors, such as PD-L1, T cell immunoglobulin and mucin domain-containing protein 3, and the tryptophan-catabolizing protein, IDO. The TME of cHL exhibits compartmental niches where these phagocytes form dense clusters around malignant cells to facilitate direct cell-cell interactions that promote immune evasion and tumor growth¹²⁹. IDO1 and tryptophan 2,3-dioxygenase (TDO2) mediate tryptophan degradation to produce kynurenine and its derivatives, which activate AhR. However, Sadik et al. showed that Interleukin 4 induced 1 is more strongly and frequently associated with AhR compared to TDO1 or TDO2¹³⁰. In addition, microbiome-produced tryptophan metabolites have recently emerged as activators of the TAM AhR. TAMs were shown to exhibit higher AhR activity compared to resident macrophages and Ahr-deficient macrophages become pro-inflammatory. Kynurenine alone was verified to be a relatively weak activator of AhR, whereas tryptophan metabolite indoles [e.g., indole, indole-3-aldehyde (IAld), indole-3-acetic acid (IAA), Indole-3-propionic acid (IPA) and Indole-3-lactic acid (ILA)] produced by the gut microbiota (specifically some Lactobacillus spp.) act as potent co-agonists with kynurenine to activate TAM AhR far more effectively than either molecule alone. These reprogrammed TAMs highly express canonical M2-like macrophage markers (CD206, IL-10, and VEGF). The reprogrammed TAMs directly suppress the activity of cytotoxic CD8⁺ T cells within the tumor, thereby promoting PDAC tumor growth⁵⁴.

Emerging evidence has demonstrated that serine, an essential amino acid for nucleotide/protein/lipid synthesis, orchestrates macrophage phenotype polarization and/or functional transition, is involved in oncogenesis and progression. The TME increases the activity of the protein kinase RNA-like endoplasmic reticulum kinase (PERK) signalling cascade in TAMs and promotes M2-like TAM activation and proliferation. PERK activation mediates serine biosynthesis via the downstream transcription factor, ATF4, a master regulator of amino acid metabolism that increases serine biosynthesis. Furthermore, increased serine enhances mitochondrial function and α-KG production, which are required for Jumonji domain containing-3 (JMJD3)-dependent epigenetic modification. Raines et al. performed an unbiased ChIP-Seq analysis of macrophages that were either PERK-sufficient or -deficient. PERK-deficient (cKO) macrophages showed greater enrichment of regions with increased H3K27 methylation in M2-like macrophages than M1-like macrophages compared to PERK wild-type (WT) macrophages. Specifically, H3K27 methylation was elevated at the loci of canonical M2 genes, including interferon regulatory factor 4 (Irf4), peroxisome proliferator-activated receptor γ (Pparg), and macrophage galactose-type lectin 2 (Mgl2), in PERK-deficient M2-like cells¹³¹. Mitochondrial dysfunction is a pivotal factor in serine-mediated macrophage education. mtROS stabilize HIF-1α in pro-inflammatory macrophages. HIF-1α is translocated to the nucleus and drives the expression of nuclear respiratory factors nuclear respiratory factor 1 (NRF1)/specificity protein 1 (Sp1). These factors activate gene transcription for mitochondrial biogenesis (e.g., Mitochondrial transcription factor a and respiratory chain components), leading to the secretion of inflammatory factors. In contrast, de novo serine synthesis hinders this pathway by promoting cytoplasmic glutathione (GSH) synthesis to deprive mtROS via phosphoglycerate dehydrogenase (PHGDH), the first enzyme in the serine synthesis pathway¹². This proposed mechanism in macrophages contrasts with the mechanism in RCC in which HIF-1 negatively regulates mitochondrial biogenesis through repression of C-MYC.

Polyamines are derived from amino acids and are essential nitrogen-containing compounds in cells. Polyamines are synthesized from L-ornithine. The precursor of L-ornithine is arginine, which is acted on by ornithine decarboxylase (ODC, also known as ODC1). Polyamine synthesis causes an increase in the putrescine level, which hinders pro-inflammatory macrophage polarization¹³². Cancer cells and other pro-tumorigenic cell types (e.g., MDSCs, TAMs, and Tregs) are highly dependent on polyamines for survival in a wide variety of cancers. The intra- and extra-cellular polyamine pools of cancer cells increase to support the fundamental cellular processes and functions of TAMs¹³³. Arginine levels are downregulated in the pro-tumor TAMs of breast cancer, whereas the levels of the polyamine products (ornithine, putrescine, spermidine, and spermine) are upregulated (upregulated Arg1 depletes arginine and produces ornithine). Mechanistically, spermine upregulates thymine DNA glycosylase (TDG) expression via triggering the p53 signalling pathway. This upregulation leads to pro-tumor TAM skewing via TDG-mediated DNA demethylation of the target gene, peroxisome-proliferator-activated receptor gamma (PPARG). PPARγ activation leads to the transcription of CD274, TGFB1, and IL-10¹⁵.

Studies have demonstrated that amino acid metabolism in B cells regulates the function of TAMs. Zhang et al. used stable isotope tracing with ¹³C₅ and ¹⁵N₂-glutamine, and confirmed that, unlike T cells, B cells can convert glutamine into gamma-aminobutyric acid (GABA) and secrete glutamine extracellularly. B-cell-derived GABA was identified as a key mediator suppressing anti-tumor immunity in a CRC mouse model. To explore the underlying mechanism, mouse or human monocytes were treated with GABA, which induced differentiation into macrophages. These macrophages exhibited higher proliferation and survival rates, increased OXPHOS capacity, and elevated expression of folate receptor β (FRβ), a marker associated with an M2-like phenotype. Furthermore, these cells secreted high levels of IL-10, which potently inhibited the cytotoxic function of CD8⁺ T cells¹³⁴. Amino acid metabolism-related signaling pathways in TAMs are summarized in Figure 4.

Bile acid synthesis occurs primarily in the liver via classical and alternative cholesterol pathways. Bile acid modification and transformation are carried out mainly by the gut microbiota in the intestine¹³⁵. Macrophages do not perform classical de novo bile acid synthesis, but macrophages can sense, take up, and respond to bile acids and metabolites within the TME, thereby modulating bile acid functions¹³⁶. Cytochrome P450 enzymes, such as CYP7A1 and CYP27A1, act as rate-limiting enzymes in bile acid synthesis. These enzymes are highly expressed in hepatocytes but are nearly absent or expressed minimal in immune cells, such as macrophages¹³⁷. Macrophages express various bile acid receptors on surfaces and intracellularly, including the nuclear receptors, farnesoid X receptor (FXR) and vitamin D receptor (VDR), as well as the membrane receptor, takeda G-protein‒receptor 5 (TGR5). For example, deoxycholic acid can act through the TGR5 signaling pathway to reduce the secretion of pro-inflammatory cytokines, such as IL-1β, IL-6, IL-12p70, and TNF-α, by macrophages, thereby exerting anti-inflammatory effects¹³⁸. Sirtuin 5 has a critical role in a Yap5SA-driven mouse model of liver cancer. Excessive succinylation of the peroxisomal enzyme, acyl-coa oxidase 2 (ACOX2), leads to activation of the bile acid synthesis pathway in Sirt5-deficient hepatocytes and increased production of taurocholic acid (T-CA). T-CA promotes the polarization of M2-like TAMs through paracrine signaling, which in turn drives liver cancer progression¹³⁹.

Iron is an essential cofactor for key proteins, such as cytochromes, in the mitochondrial electron transport chain (ETC). Tumor cells typically exhibit extremely high iron requirements to support proliferation, DNA synthesis, and energy metabolism¹⁴⁰. To meet this demand, tumor cells hijack iron by upregulating transferrin receptor 1 (TFR1), creating a relatively iron-deficient and hypoxic state within the TME, particularly in the tumor core. Decreased intracellular iron levels or damage to iron-sulfur clusters can impair the antibacterial and anti-tumor functions of M1 macrophages. Cancer cells compete with macrophages in HCC for iron by overexpressing transferrin receptor (TFRC). Analyses of human and mouse HCC samples have shown that ferrous ion levels are significantly lower in M2-like TAMs than M1 counterparts. Furthermore, treating macrophages with iron chelators, such as deferasirox (DFX) or ciclopirox (CPX), upregulates the levels of M2 marker gene expression (e.g., Arg1, Vegfa, and PD-L1) and suppresses T cell proliferation⁵⁹. CD63⁺ macrophage subsets constitute a key immunomodulatory node in the bone metastasis microenvironment with the immunosuppressive function being directly attributable to aberrant iron metabolism. Therapeutic application of a CD63-targeted nanosystem loaded with a ferroptosis inducer in a mouse model of bone metastasis promoted the infiltration of M1-like TAMs and cytotoxic CD8⁺ T cells into the TME and significantly reduced the bone metastasis burden¹⁴¹.

The core mechanism involves TAMs being “domesticated” by tumor cells to act as accomplices in promoting cancer. Extensive research has revealed that a high abundance of TAMs is strongly linked to poor patient prognosis across many cancer types¹⁴². Thus, strategies to reduce TAM infiltration in the TME remain among the most extensively studied approaches. For example, blocking CSF-1R, which is primarily expressed on TAMs, suppresses TAM survival. Inhibiting the CCL2/CCR2 axis directly reduces monocyte recruitment to the TME. Furthermore, clodronate liposomes, including the novel Clo-Lipo-DOTAP formulation, have been shown to effectively deplete TAMs by inducing apoptosis via phagocytosis in melanoma and colon cancer models¹⁴³. This approach further validates the strategy of TAM depletion, although it has thus far been applied primarily in preclinical mouse models. Current monotherapy approaches aimed at reducing the number of TAMs face significant limitations. These limitations include side effects, such as neutropenia and thrombocytopenia, as well as low response rates in solid tumors. Chen et al. demonstrated that the CSF1R inhibitors (BLZ945 and PLX3397) alone had only a modest anti-tumor effect on head and neck squamous cell carcinoma (HNSCC). However, when combined with cisplatin, CSF1R inhibitors significantly suppressed tumor progression compared to monotherapy¹⁴⁴. Targeting the CCL2/CCR2 axis is another approach to reduce monocyte/macrophage numbers in the TME. However, this approach typically induces only a transient reduction because the approach is counteracted by a compensatory increase in CXCL1/3. Given these limitations, combination therapies have shown superior efficacy to monotherapy. These strategies extend beyond combinations with anti-PD-1¹⁴⁵. For instance, a recent study used a vascularized tumor tissue model to evaluate a multispecific antibody targeting CCR2, CSF-1R, and TGF-β, which successfully reduced monocyte levels¹⁴⁶. These findings indicated that an approach to deplete TAMs holds promise as part of combination therapy for specific cancers.

A more promising approach to circumvent the side effects associated with depleting monocytic TAMs is to repolarize TAMs within the TME, thereby improving therapeutic efficacy. CD40 signaling initiates metabolic reprogramming characterized by enhanced FAO and glutamine metabolism. These pathways provide substrates for ACLY, culminating in ACLY-dependent epigenetic remodeling. CD40 agonists effectively repolarize immunosuppressive M2-like TAMs toward an immunostimulatory M1-like state through this mechanism. This transition enhances antigen presentation and promotes the infiltration of CD8⁺ T cells into the TME¹²³. Nevertheless, the efficacy of CD40 agonism as monotherapy is limited and systemic activation often leads to significant toxicity. To address this finding, strategies using engineered antibodies and refined administration methods are being developed to enhance targeting of tumor or TAM¹⁴⁷. Concurrently, combination therapy with immune checkpoint inhibitors has emerged as the most common clinical strategy to maximize benefit. TREM2, a core hub for lipid metabolism regulation, has emerged as a promising new target for reprogramming the TME through TAMs with targets, such as CD40. TREM2 engages with multiple ligands, activating context-dependent signaling pathways that influence disease progression¹⁴⁸. Inhibiting TREM2 to disrupt immunosuppressive networks has shown therapeutic potential in oncology, albeit with outcomes that vary across cancer types. Monoclonal antibodies and chimeric antigen receptor macrophages (CAR-Ms) are currently under investigation in preclinical and early clinical development. The Amit laboratory (Weizmann Institute of Science, Rehovot, Israel) recently developed myeloid-targeted immunocytokines and NK/T-cell enhancers (MiTEs), a novel antibody-based therapy that depletes TREM2⁺ TAMs and acts in a tumor-restricted manner to enhance killer cell cytotoxicity¹⁴⁹. This finding represents a remarkable discovery in which immune cells are reprogrammed to kill tumors.

The broad distribution of macrophages complicates TAMs targeting because there is a risk of off-target effects and systemic immune dysregulation. Elevated IRG1 expression in TAMs triggers M2-like polarization, making IRG1 a promising cancer target. Recently, the Yang laboratory (Huazhong University of Science and Technology, Wuhan, China) designed a lipid nanoparticle (LNP) encapsulating siRNA to effectively silence Irg1 in TAMs¹⁵⁰. This strategy demonstrated potent anti-tumor efficacy and favorable safety profiles in melanoma and colon cancer models. The mechanisms by which succinate modulates TAM polarization have been characterized and SMPs exhibit improved anticancer immunotherapy effects, establishing succinate as a viable oncologic target. A central challenge in leveraging this target is the efficient intracellular delivery of succinate into macrophages, circumventing an interaction with the surface receptor, SUCNR1. To overcome this membrane barrier, the Tang laboratory (Huazhong University of Science and Technology, Wuhan, China) developed novel microparticles to deliver high concentrations of succinate directly into the cellular interior³⁷. Furthermore, macrophage-specific folate receptor β (FR-β, encoded by folr2) is a validated cancer target that is highly expressed on TAMs. The Wang laboratory (Hebei Medical University, Shijiazhuang, China) has developed an FR-β–targeting chimeric molecule that repolarizes TAMs to an M1-like state in the TME. By linking the TLR7/8 agonist, IMDQ (a reprogramming agent lacking specificity) to a 6-substituted pyrrolo[2,3-d]pyrimidine, the construct achieves higher affinity for FR-β than folate does and demonstrates applicability across multiple cancer types¹⁵¹.

The core of therapeutic strategies targeting CD36 on TAMs involves disrupting reprogrammed lipid metabolism. This intervention aimed to reverse the protumor M2-like polarization and alleviated suppression of anti-tumor immunity. One of the earliest and most extensively studied CD36 inhibitors is sulfonyl-N-succinimide oleate (SSO). For example, Xu et al.¹⁵² demonstrated that oxidized low-density lipoprotein (oxLDL) activates p38 signaling to suppress IFN-I production in mouse and human macrophages. Blocking CD36 with SSO inhibits this process and has been shown to be effective in melanoma and CRC mouse models. Recent studies have indicated that CD36 expression is elevated in metastasis-associated macrophages isolated from metastatic liver tumors. Sheue-Fen et al. developed a humanized anti-CD36 monoclonal antibody (PLT012). By targeting metabolic pathways, PLT012 can remodel the immune microenvironment in HCC and liver metastases, affecting TAMs, CD8⁺ T cells, and NK cells, and thereby restoring anti-tumor immunity. This approach holds potential for future immunotherapy in liver cancer¹⁵³. In summary, most anti-tumor strategies directed at CD36 remain in preclinical or early-stage clinical (phase I/II) development, with no therapies formally approved for clinical use. Importantly, CD36 is also expressed on various normal cell types, including endothelial cells, platelets, and muscle cells. Consequently, potential systemic side effects, such as an impact on angiogenesis and systemic energy metabolism, as well as off-target effects, require careful consideration. In addition, inhibiting CD36 expression may drive TAMs to acquire fatty acids through alternative pathways, potentially leading to drug resistance. Addressing this adaptive response remains a key challenge in the development of CD36-targeted therapies.

Therapeutic small molecules targeting metabolic enzymes have been shown to alter the TAM phenotype and tumor progression. For example, energy deprivation can be achieved with the GLUT1 inhibitor, WZB117, whereas PD-L1 expression is downregulated by the PFKFB3 inhibitor, 3PO. Inhibiting LDHA with oxamate suppresses OXPHOS and promotes repolarization toward an M1-like phenotype. Reprogramming lipid metabolism via the LXR agonist, GW3965, or the FABP5 inhibitor, SBFI-26, reduces the population of protumor TAMs. Similarly, research has confirmed the effectiveness of the glutaminase inhibitor, JHU083, and the Arg1 inhibitor, CB-1158¹⁵⁴. Most of these small molecules have exhibited potential in preclinical studies. However, translation to clinical success is not guaranteed. A case in point is the IDO inhibitor, epacadostat, which demonstrated excellent anticancer efficacy in earlier research but ultimately met with setbacks in a phase III clinical trial. Moreover, the development of these metabolic immunotherapy approaches is significantly hampered by the persistent challenge of effectively targeting TAMs.

The phagocytosis of tumor antigens and direct clearance of cancer cells are critical functions of macrophages. However, this activity is highly suppressed within the TME and even systemically suppressed due to the abundance of immunosuppressive signals. The interaction between inhibitory ligands expressed by tumor cells and receptors on TAMs activates suppressive intracellular signals. This interaction leads to the repression of phagocytosis and promotion of immune evasion. Among these intracellular signsl, “do not eat me” signals are the most significant inhibitory component. Preclinical and early-stage clinical trials have indicated that blocking these inhibitory phagocytic checkpoints or restoring “eat me” signals represents a promising therapeutic strategy for cancer.

The CD47–SIRPα signaling axis is a dominant “do not eat me” signaling pathway that protects cells, including cancer cells, from phagocytic clearance. Early anti-CD47 therapies induce severe blood toxicity because CD47 is ubiquitously expressed on cancer cells and red blood cells (RBCs). Current next-generation agents, such as the Fc-fusion-mutant SIRPα protein, HCB101, are engineered to bind CD47 on cancer cells with high affinity, while exhibiting minimal binding to RBCs¹⁵⁵. A major challenge in therapeutic strategies targeting the CD47–SIRPα checkpoint is the compensatory triggering of the CD24–Siglec-10 axis, another prominent “do not eat me” signal that contributes to cancer immune evasion. Anti-CD24 and the anti-CD24/anti-CD47 combination induced higher phagocytic activity against ovarian cancer cells than anti-CD47 alone. Thus, targeting the CD24–Siglec10 axis has become a rapidly developing strategy and several monoclonal anti-CD24 antibodies are currently in clinical trials, e.g., PHST001, IMM47 and ATG-031 (ClinicalTrials.gov). In addition, Zhang et al. developed an implantable microneedle device for surgical cavity placement to prevent postoperative tumor recurrence. The microneedle delivers macrophage-activating nanoparticles (GMANs), which block the CD24–Siglec10 “do not eat me” signaling pathway, thereby enhancing macrophage recognition and phagocytosis of tumor cells¹⁵⁶. Concurrently, the GMANs inhibit the expression of oncostatin M (OSM), helping to reverse postoperative immunosuppression within the TME. Furthermore, the microneedle is conjugated with the immune adjuvant CpG oligonucleotide to recruit immune cells to the site and generate a localized anti-tumor immune response. Although CD24–Siglec10 blockade has a superior blood safety profile compared to CD47 blockade, studies have indicated that CD24–Siglec10 blockade may compensate for activation of the CD47–SIRPα pathway. Future therapies may combine these monoclonal antibodies with CAR-M and/or other targeted agents.

Tumor cells can achieve immune evasion by overexpressing glycans modified with sialic acid (SA). Siglecs, the primary receptors for SA, are expressed predominantly on immune cell surfaces. Binding to SA ligands transmits positive or negative regulatory signals into cells, thereby actively modulating innate and adaptive immune responses¹⁵⁷. Therapeutically, several monoclonal antibodies have been developed to target Siglecs. For example, gemtuzumab ozogamicin (Mylotarg), which targets CD33 (Siglec-3), is already approved for use in acute myeloid leukemia (AML). In addition, NC318, an antibody targeting Siglec-15, is currently undergoing clinical trials¹⁵⁸. Another innovative approach involves enzymatic desialylation. E-602 is a novel, first-in-class engineered human sialidase (Neu2) Fc fusion protein designed to enhance anti-tumor immunity by removing immunosuppressive sialic acid residues from glycans, which is now under development for cancer treatment. Furthermore, a dual-function PD-L1-targeting sialidase fusion molecule has entered clinical trials. The current immune therapeutics to modulate TAMs are list in Table 1.