Targeting macrophages in cancer immunotherapy: Frontiers and challenges

Graphical abstract

Highlights

• •The alteration of the M1 and M2 macrophage phenotype is regulated by various factors.
• •Individualized CAR-macrophage (CAR-M) therapy has specific advantages and will be a promising cancer immunotherapy.
• •Targeting phagocytosis by inhibiting checkpoints or their receptors or blocking their interaction is a novel approach to cancer immunotherapy.
• •Clinical trials for evaluating treatments that target phagocytosis checkpoints show their promise in cancer immunotherapy.

Abstract

Background

Cancer immunotherapy has emerged as a groundbreaking approach in cancer treatment, primarily realized through the manipulation of immune cells, notably T cell adoption and immune checkpoint blockade. Nevertheless, the manipulation of T cells encounters formidable hurdles. Macrophages, serving as the pivotal link between innate and adaptive immunity, play crucial roles in phagocytosis, cytokine secretion, and antigen presentation. Consequently, macrophage-targeted therapies have garnered significant attention.

Aim of review

We aim to provide the most cutting-edge insights and future perspectives for macrophage-targeted therapies, fostering the development of novel and effective cancer treatments.

Key scientific concepts of review

To date, the forefront strategies for macrophage targeting encompass: altering their plasticity, harnessing CAR-macrophages, and targeting phagocytosis checkpoints. Macrophages are characterized by their remarkable diversity and plasticity, offering a unique therapeutic target. In this context, we critically analyze the innovative strategies aimed at transforming macrophages from their M2 (tumor-promoting) to M1 (tumor-suppressing) phenotype. Furthermore, we delve into the design principles, developmental progress, and advantages of CAR-macrophages. Additionally, we illuminate the challenges encountered in targeting phagocytosis checkpoints on macrophages and propose potential strategies to overcome these obstacles.

Introduction

Cancer immunotherapy eliminates cancer cells and restricts metastasis and recurrence by mobilizing the human body’s own powerful immune system, it is mainly realized by the manipulation of immune cells or the immune checkpoint blockade. The engineered immune cells enhance the immune system function and prolong the survival time of tumor patients. For example, the chimeric antigen receptor (CAR)-T immunotherapy approved by FDA has achieved success in the treatment of hematological malignancies [1], [2]. Therefore, cancer immunotherapy is highly suggested due to its higher safety and fewer side effects, it has achieved a breakthrough in both basic research and clinical application. However, a couple of formidable problems including insufficient CAR-T cell production, meager collection of T cells and soaring cancer progress result in the failure of CAR-T therapy. Therefore, new strategies on cancer immunotherapies need to be developed urgently. The recently developed CAR-Natural Killer (NK) cell and CAR-Macrophage (CAR-M) are showing their promising result in mice models of cancer immunotherapy. Compared with CAR-T and CAR-NK, CAR-Macrophages have specific advantages such as low toxicity and high efficacy.

Macrophages are heterogeneous and multifunctional cells with high plasticity, they are primarily involved in pathogens elimination, cellular debris clearance, inflammatory response regulation and cell homeostasis [3], [4]. Targeting macrophages has several advantages in tumor immunotherapy. First, low invasion is the main obstacle to T cell-based anti-cancer treatment, but macrophages occupy 30–50 % among all the infiltrating immune cells in the tumor microenvironment [5]. Second, the quantity of circulating monocytes is in the highest proportion of infiltrating macrophages in cancers, and the availability of peripheral blood monocytes supports feasibility and implementation of macrophage-based treatment strategies in clinical practice.

Macrophage-targeting strategies include cytokines or chemokine inhibition that promote the recruitment and polarization of myeloid cells, as well as activators of their anti-tumor and immune stimulation functions. Macrophages can be induced into M1 (anti-tumor) or M2 (pro-tumor) phenotype and novel strategies have been developed to alter macrophages from M2 to M1 phenotype. The polarization were orchestrated by lots of factors including tumor microenvironment, oncogenes, tumor suppressors, non-coding RNA, small molecules and others. Targeting these components will bring new therapeutic windows in cancer treatment. In addition, early clinical trials have shown that immune checkpoints targeting myeloid cell function have promising anti-tumor potential. More important, CAR-M is being evaluated in clinical trials and current pre-clinical trials results also suggest macrophages are excellent candidates for cell therapy.

Regarding to the immune checkpoint blockade, the adaptive immune checkpoints such as cytotoxic T lymphocyte-associated antigen-4, (CTLA-4) and programmed cell death ligand 1 (PD-L1) have been well studied [6], [7], [8], inhibition of T-cell checkpoints activates effector T cells by blocking the T cell inhibition pathway and enhancing the anti-tumor immune response. However, the low response rate of solid tumors drags our attention to the innate immune cells. Besides antigen presentation, dendritic cells and macrophages also serve as the phagocytes in engulfment of cancer cells. Tumor cells present “don’t eat me” signals that could contribute as phagocytosis checkpoints by interacting with receptors on the phagocytes. After the cluster of differentiation 47 (CD47) was identified as a phagocytosis checkpoint, more phagocytosis checkpoints were discovered in a recent couple of years (Fig. 1), including PD-L1 [9],major histocompatibility class I complex (MHC-I),the cluster of differentiation 24 (CD24), CD22 [12], stanniocalcin 1(STC1) [13], GD2 [14]. Many clinical trials on targeting phagocytosis checkpoints are under processing, but no drugs in the market now.

• (A)The phagocytosis process of macrophage engulfment of cancer cells. (B) The emerging phagocytosis checkpoints in recent years were studied in both basic research and clinical trials. (1) CD47 expressed on cancer cells interacts with SIRPα on phagocytes including monocytes, macrophages and DCs inhibits the phagocytosis of tumor cells by phagocytes. (2) CD24 expressed on the surface of the cancer cells binds to Siglec-10 in the phagocyte and inhibits phagocytosis of cancer cells by phagocytes. (3) The interaction between PD-1-PD-L1 not only suppresses T cell immune response but also inhibits phagocytosis of cancer cells by macrophages. (4) The high expression of STC1 in cancer cell interacts with calreticulin (“eat me” signal) and suppress the membrane calreticulin-derived phagocytosis by phagocyte and inhibit antigen presentation from APCs to T cells. (5) The β2M subunit of MHC-I on cancer cell binds to LILRB in phagocyte and inhibit the innate immune response, thus the phagocytosis of cancer cells by macrophages was blocked. (6) The disialoganglioside GD2 highly enriched in brain tumors binds to Siglec-7 on macrophages and inhibits the phagocytosis of neuroblastoma cancer cells by macrophages. (7) CD22 is a cell surface sialoglycoprotein exclusively expressed on B cells. In B cell malignancy, it binds to α2,6-linked sialic acid on microglia and inhibits the phagocytosis of B cells.

Fig. 1.

Phagocytosis checkpoints in cancer immunotherapy.

In a word, the macrophage-mediated immunomodulation in cancer immunotherapy has emerged as a prosperous therapeutic strategy and is expected to the next frontier in tumor immunotherapy [15], [16]. However, even different strategies are developing fast, there is no clinical trials for the macrophage polarization application and very few clinical trials on CAR-macrophages, the main obstacles and restrictions will be discussed. Targeting phagocytosis checkpoints are under the most advanced development and clinical trials are in full swing. We will discuss the challenges and provide suggestions on solutions in this review. Here, we mainly emphasize three new frontiers: polarization rewiring, CAR-macrophage, and targeting phagocytosis checkpoints.

Orchestration of macrophage polarization in cancer immunotherapy

Concept of Macrophage polarization

Macrophages are highly plastic cells that undergo different forms of activation of different signals [17], they have polarized into two subsets under a special microenvironment: classically activated M1 macrophages and alternatively activated M2 macrophages which have 4 subcategories: M2a, M2b, M2c and M2d. In cancer, the signals that coordinate this plasticity vary greatly between different tumors or between different parts and stages of the same tumor, leading to different tumor associated macrophage (TAM) phenotypes. Both phenotypes have different cell metabolism, cytokine secretion and immune functions and they can switch from one phenotype to the other under specific stimulation.

Regarding cell metabolism, the M1 type macrophages are involved in metabolic reprogramming toward glycolysis [18], while the M2 macrophage rich with lipids use fatty acid oxidation (FAO) rather than glycolysis to obtain their energy resource, and FAO facilitates IL-1β secretion to contribute to macrophage polarization to M2 phenotype [19]. TAMs with high FAO increased oxidative phosphorylation, dephosphorylation of SHP1 and phosphorylation of JAK1, which regulates the polarization of macrophages [20]. The induction of M1 and M2 phenotypes are listed in Fig. 2. Markers of M1 such as IL-6, IL-1β, iNOS and markers of M2 such as IL-10, CD163, CD206, Arg1, IL-13 are often used to distinguish M1 and M2.

Fig. 2.

Factors orchestrate macrophage polarization.

The molecular rewiring of macrophage polarization

The nomenclature of M1 and M2 were brought by Charlie Mills in 2000 [21]. Macrophages were named M1 or M2 macrophages according to their activation status, functions and different cytokine secretion. There are four subtypes of M2 macrophages (M2a, M2b, M2c and M2d) and each responses to different activators and exhibits different biomarkers and functions. The stimuli and functions of each type were listed in Table 1. M1 macrophage exhibits anticancer activity by releasing nitric oxide and stimulating T cells to produce Th1 cytotoxic response. They are less potent in antigen presentation and phagocytosis compared to the pro-tumoral M2 macrophages [22]. While M2 macrophages secret heterogeneous cytokines to prompt tumor growth. Therefore, the polarization of macrophage from M2 type to M1 type is a promising strategy to improve T-cell mediated anti-tumor immunity and improve the immunosuppressive tumor microenvironment. Tumor-associated macrophages (TAM) infiltrated in the tumor microenvironment exhibited the phenotype and function similar to M2 macrophages and play essential roles in cancer immunotherapy. TAM constitute about half of the tumor mass, indicating their highest infiltrations among all the immune cells including T cells and NK cells. TAM directly reduce the activity of T cells and NKs by expressing cell surface proteins or releasing soluble factors with immunosuppressive function, or recruiting other immunosuppressive cells to evade local immune surveillance. TAMs in adoptive cell transfer therapy support tumor angiogenesis and fibrosis TME by producing pro-angiogenic growth factors [23], hindering the infiltration of adoptive cells into the tumor site. In addition, high TAM penetration is associated with a low expression of IFN-γ active NKs in tumor [24], which may have a negative impact on the activation or survival of NKs in adoptive cell metastasis.

Table 1.

The classical stimuli and main functions of M1 and M2 subtype macrophages.

Types	Activating stimuli	Biomarkers	Main functions
M1	LPS + INFγ	CD86,CD80,CD68,MHC-II,TLR-2, TLR-4, IL-1R, IL-10low, IL-12high, iNOS	1) Immunosuppression 2) Th1-dependent inflammation cells activation 3) Th-1 helper cells activation 4) Intracellular pathogens destruction
	GM-CSF
	TNF-α
M2a	IL-4	Human: CD206, IL1Ra:IL-1RII Mouse: Arg-1, FIZZ1, Ym1/2	1) Destruction of parasites 2) Allergic reactions 3) Th2-depended inflammation
	IL-13
	IRF4
	Imjd3
M2b	IL-1β	IL-10 high, IL-12 low, CD86	1) Th2 helper cells activation 2) Pro-inflammatory induction, increase the release of IL-10, IL-1β, Il-6.
	Immunoglobulin complexes
M2c	IL-10	Human:CD206, TLR-1, TLR-8; mouse: Arg-1	Regeneration of immunosuppression Anti-inflammatory phenotype to apoptosis cells
	TGFβ
	Glucocorticoids
M2d	IL-10	VEGF, IL-12 low, TNF-α low, IL-10 high	Potentiation on angiogenesis and tumor progression
	TGFβ
	A2AR

The imbalance of macrophage M1-M2 polarization is usually associated with various immune mediated-diseases and inflammatory conditions. Generally, in the early stage of a tumor, macrophages are more inclined to exhibit a positive anti-tumor immune response, while in the late stage of cancers, they are more seduced by tumor cells and prone to pro-tumor cells. However, many macrophages exhibit mixed characteristics of M1 and M2 or are in the process of polarization. The alteration of M1 and M2 is regulated by various factors. Here, we mainly summarize the novel factors discovered in recent years (Fig. 2, Table 2).

Table 2.

Regulators of macrophage polarization.

Regulators	Macrophage polarization	Working mechanisms	Literature
Hypoxia	From M1 to M2		[26], [27], [28]
Hypoxic tumor-derived exosomal miR-301a	Promoted M2 polarization	By PTEN/PI3Kγ to promote cancer progression	[26]
The hypoxic glioma-derived exosomes	Promoted M2 polarization	IL-6-pSTAT3-miR-155-3p-autophagy-pSTAT3 positive feedback loop	[29]
Lung tumor extracellular vesicles	Promoted M2 polarization	extracellular vesicles with miR-19b-3p promoted lung cancer metastasis by Hippo signaling	[33]
Extracellular vesicles	Induced macrophage M2 polarization	activating AKT-STAT3/6 pathway and down-regulating PTEN expression	[34]
Tumor-derived exosomal miR-934	Induced macrophage M2 polarization	inhibiting PTEN and increasing the PI3K/AKT signaling	[35]
MSC-EVs	Promoted M2 polarization	anti-inflammatory and immunosuppressive properties	[36], [37]
CSF-1/CSF-1R	Promoted M2 polarization	CSF-1 and IFN-γ promotes M2 markers expression	[39]
Succinate	Promoted M2 polarization	SUNCR1- activated PI3K-hypoxia-inducible factor 11α (HIF-1α) signaling	[41]
Lactate	Promoted M2 polarization	through MCT-HIF1α	[42], [43]
α-ketoglutarate	Augmented M2 macrophage polarization	α-ketoglutarate accumulation led by the activation of SENP1-Sirt3 signaling	[45]
Itaconic Acids	Promoted M2 polarization	IRG1 deficiency exhibit pro-inflammatory features to inhibit tumor growth	[46]
Serine	Promoted M1 polarization	increasing IGF1 through the activation of p38-JAK-STAT1pathway	[48]
One-carbon metabolism	Induced M1 polarization	increasing S-adenosylmethionine level and histone methylation	[49]
Glucosylceramide	Sustained M1 polarization	induced IRE1-mediated spliced XBP1 production and STAT3 activation	[50]
GABA	Promoted M2 polarization	secrets IL-10 and suppresses cytotoxic CD8+T cells	[51]
Itaconate	Inhibited M2 polarization	Inhibiting phosphorylation of JAK1 and STAT6	[52]
Luteolin	Induced macrophage M2 polarization	downregulating p-STAT3 and upregulating p-STAT6 [53]	[53]
Physalin D	From M1 to M2	Suppressing STAT1 and blocking its nuclear translocation	[55]
Nanoparticles from traditional Chinese plant Panax Ginseng	Induced M2 to M1	Increased ROS and suppressed tumor growth	[56]
Metformin	Induced M2 to M1	Recruitment of CD8+ cells and decrease of regulatory T cells	[58]
Etomoxir	Inhibited polarization	depleting intracellular free coenzyme A (CoA)	[59]
Pirfenidone	Reducing M2 macrophage	reducing M2 macrophage infiltration and inhibiting TGFβ1-Smad 3 signaling activation	[60]
Vitamin D	Inhibited M1 polarization	STAT-1/TREM-1 pathway	[61]
Decitabine	Promoted M2 polarization	Demethylation of PPARγ	[62]
Trametinib	Inhibited M2 polarization	MEK inhibitor	[63]
Panobinostat	Inhibited M2 polarization	HDAC inhibitor	[63]
Ultrasound-driven discharge of piezoelectric materials	Promoted M1 polarization	enhanced the release of the proinflammatory factors TNF-α and IL-1β	[64]
Aryl hydrocarbon receptor	Suppressed M1 polarization	via the AhR-miR-142a-IRF1/HIF-1αpathway.	[67]
High-mobility gene group A2 (HMGA2)	promoted M2 polarization	activating STAT3 and inducing CCL2 secretion	[68]
Krüppel-like factor 4 (KLF4)	Induced M2 polarization	interacts with STAT6	[69]
Myocyte enhancer factor 2 C (MEF2C)	Promoted M1 polarization	Promotes interleukin-12 p35 subunit (Il12a) and interleukin-12 p40 subunit (Il12b) thus created a pro-inflammatory environment	[70]
Progranulin	Inhibited M1 polarization	down regulating NF-кB pathways	[72]
Annexin A5	Promoted M2 polarization	directly interacting with PKM2 at ASP101, LEU104 and ARG106	[73]
Annexin A1	Promoted M2 polarization	AMPK-mTOR pathway [74]	[74]
METTL3	Potentiate M1 polarization	Mettl3 deficiency impairs the TLR4 signaling	[80], [81]
MTHFD2	Promoted M2 polarization	regulating MTHFD2-PTEN interaction	[75]
Acetylation of STAT6	Suppressed M2 polarization	Stat6 is acetylated by CBP, Trim24, a CBP-associated E3 ligase, promotes Stat6 acetylation by catalyzing CBP ubiquitination	[76]
miR-16	Promoted to M1 polarization	Inhibits TAM infiltration	[82]
miR-19a-3p	Induced M2 polarization	regulating TNFα-induced protein 3-NF-kB feedback loops	[83]
Non-coding RNA (lncRNA) ANCR	Inhibited macrophage M1 polarization	reduced the level of IL-1β and IL-6	[84]
lncRNA BCRT1	Potentiated the M2 macrophage	by exosome mediated migration	[85]
lncRNA COX2	Enhanced inflammation and M1 macrophage phenotype	Regulation of CREB-C/EBPβ pathway	[87]
LncRNA-MM2P	Enhanced M1 and blocking M2 polarization	suppressing STAT6 phosphorylation	[88]

Macrophages can be induced into M1 or M2 phenotypes. Macrophages in M1 phenotype exhibit anti-tumor and pro-inflammatory features. The cytokines such as IL-4, IL-10 and IL-13 treatment induce macrophages into M1 phenotype. The treatment of glucocorticoids and glucosylceramide induces macrophages into M1 phenotype. Furthermore, the activation of one-carbon metabolism and NK, T cells or the deficiency of serine metabolism induce macrophage into M1 phenotype. The nanoparticles derived from traditional Chinese plant Panax Ginseng induced M2 macrophage to M1 macrophage, increased the accumulation of reactive oxygen species (ROS) and suppressed melanoma tumor growth. Luteolin treatment alters macrophages from M1 proinflammatory type to M2 anti-inflammatory type via downregulating p-STAT3 and upregulating p-STAT6.^. METTL3 as an RNA methyltransferase potentiate the M1 macrophage polarization, deficiency of METTL3 suppressed TLR signaling-mediated macrophage activation and increased the M2 macrophages in tumor microenvironment. Macrophages can be induced into M2 phenotypes by treatment of HhR, HMGA2, GABA, succinate, lactate and α-ketoglutarate, it can also be induced by miRNAs such as miR-934. Macrophages can be orchestrated into M1 or M2 phenotypes. For example, the luteolin treatment, histone lactylation alter M1 to M2, while the metformin treatment and the IRG1/ITA deficiency alter M2 to M1. Various CAR-macrophage such as Adenovirus contained CAR-macrophage alter the M2 macrophage to M1 macrophage. MEF2C promoted M1 polarization and the expression of interleukin-12 p35 subunit (Il12a) and interleukin-12 p40 subunit (Il12b) thus created a pro-inflammatory environment.

(MCH: monocarboxylate channel transporter, SUCNR1: Succinate Receptor 1, GC:glucosylceramide PI3K: phosphatidylinositol 3-kinase, NF-kB: nuclear factor kB, HIF-1α: hypoxia inducible factor-α, PPARγ: peroxisome proliferater- activated receptor, SAM: S-adenosyl methionine, JAK: Janus kinases, LPS: Lipopolysaccharide, METTL3: methyltransferase like 3, KLF4: Kruppel-Like Factor 4, PTEN: phosphatase and tensin homolog deleted on chromosome ten, PKM2: Pyruvate kinase M2 subtype, IFNγ: Interferon gamma.)

The tumor microenvironment orchestrates macrophage polarization

Tumor microenvironment is mixed with components of immune cells, tumor cells, normal cells and other exosomes, it determines cancer progress and influences the cancer immunotherapy and outcomes [25]. Macrophage as one of the most important innate immune cells, its polarization and function are regulated by tumor microenvironment at different levels.

First, O₂ saturation orchestrates the energy supply in the tumor microenvironment, hypoxia promoted cell migration to the center of the primary tumor, it facilitated macrophages to polarize into M2 phenotype instead of M1 phenotype, this has been demonstrated in many cell and in vivo tumor models [26], [27], [28]. In pancreatic cancer, the hypoxic tumor-derived exosomal miR-301a mediated M2 macrophage polarization via PTEN/PI3Kγ to promote cancer progression [26]. In glioma, the hypoxic glioma-derived exosomes (HGDEs) promoted macrophage M2 polarization and thus potentiated the glioma progression in vivo, the HGDEs highly expressed IL-6 and miR-155-3p and induced M2 polarization by the IL-6-pSTAT3-miR-155-3p-autophagy-pSTAT3 positive feedback loop [29]. However, in a recent study, it is indicated that the chronic physiological hypoxia enhanced the efferocytosis ability of macrophages [30], even it didn’t mention the macrophage polarization, but higher phagocytosis ability usually is a typical characteristics of M1 macrophage. Therefore, the hypoxia extent and speed probably also affect macrophage polarization and it is a complicated process required further investigation.

The extracellular vesicles including exosomes released by cells are important components in tumor microenvironment, they modulate immune cell functions especially involve in macrophage biological behaviors in cancer development and cancer immunotherapy [31], [32]. In lung cancer, the M2 polarization together with the tumor derived extracellular vesicles with miR-19b-3p promoted lung cancer metastasis by Hippo signaling [33]; in breast cancer, the extracellular vesicles derived from breast cancer induced macrophage M2 polarization by activating AKT-STAT3/6 pathway and down-regulating PTEN expression, and thus potentiating cancer progression [34]. Moreover, the exosomes influence macrophage polarization. The tumor-derived exosomal miR-934 induces the macrophages to M2 polarization by inhibiting PTEN and increasing the PI3K/AKT signaling to enhance colon cancer metastasis [35]. In addition, mesenchymal stem cell-derived extracellular vesicles (MSC-EVs)-based therapy has been investigated a lot, the MSC-derived exosomes promotes M2 polarization with anti-inflammatory and immunosuppressive properties [36], [37].

The immune cells, chemokines and other immune components also contribute to macrophage polarization. For example, the activation of Natural Killer T cells increases M1 macrophages and effector Th1 cells in tumor microenvironment and inhibits tumor growth [38]. More importantly, the classical stimulator for macrophage polarization IFN-γ and interleukin 4 (IL-4) are released by T cells. Macrophages are activated by the immune response with the cytokines secretion to phagocyte tumor cells. In addition, the colony stimulating factor-1 (CSF-1) contributes to the differentiation and survival of macrophages, inhibition of CSF-1 or its receptor CSF-1R decreased the M2 markers of macrophages and suppressed glioma growth [39].

Taken together, the extrinsic polarization of macrophages modulated by various factors in the tumor microenvironment contribute to tumor progression.

The cellular metabolites affect macrophage polarization

The homeostasis of macrophage were modulated by metabolites, nutrition, growth factors, etc. [40]. Succinates released from cancer cells to the microenvironment polarize macrophages to the M2 phenotype, which correspondingly promoted tumor metastasis. This process was mediated by the succinate receptor SUNCR1- activated PI3K-hypoxia-inducible factor 11α (HIF-1α) signaling [41]. Lactate accumulation induced M2-like macrophages, which improved muscle reperfusion and regeneration during ischemia and promotes tumor progression in gastric cancer through MCT-HIF1α [42], [43]; it also induced M2 polarization Mettl3/Trib1/Erk/Stat3 pathway in endometriosis [44]. The α-ketoglutarate accumulation led by the activation of SENP1-Sirt3 signaling augments M2 macrophage polarization [45]. IRG1 encoded an enzyme called ACOD1 which catalyzed the production of oncogenic Itaconic Acids (ITA), IRG1 was highly expressed in macrophages and macrophages deficient with IRG1 exhibit pro-inflammatory features to inhibit tumor growth [46]. In addition, serine was required for lipopolysaccharide (LPS) induction of IL-1β instead of inflammation activation [47], serine metabolism deficiency promoted M1 polarization and inhibits STAT6-mediated M2 polarization by increasing IGF1 through the activation of p38-JAK-STAT1 pathway [48]. One-carbon metabolism also induces M1 polarization by increasing S-adenosylmethionine level and histone methylation [49]. The glucosylceramide from tumor cells induced the reshuffling of lipid composition on the endoplasmic reticulum (ER) and sustains the macrophage M1 polarization [50]. A newly emerging factor neurotransmitters γ-Aminobutyric acid (GABA) secreted by activated B lymphocytes, promoted monocyte differentiation into M2 phenotype that secreted IL-10 and suppresses cytotoxic CD8⁺T cells [51]. The TCA cycle metabolite itaconate and 4-octyl itaconate (OI) suppressed M2 macrophage polarization by inhibiting JAK1 phosphorylation STAT6 phosphorylation in vivo [52].

The natural compounds and small molecules modulate macrophage polarization

The natural compounds extracted from plants modulated macrophage polarization. For example, Luteolin, a natural flavonoid compound with potent anti-inflammatory activity. Luteolin treatment altered macrophages from M1 proinflammatory type to M2 anti-inflammatory type via downregulating p-STAT3 and upregulating p-STAT6 [53]. Physalin D was a natural phytocompounds isolated from Physalis species (Solanaceae) used for against liver fibrosis [54]. It regulated macrophage polarization by suppressing STAT1 activation and blocking its translocation to nuclear [55]. Physalin D repolarized M1 phenotype cells toward M2 phenotype and further maintained M2 phenotype [55]. Besides, the nanoparticles derived from traditional Chinese plant Panax Ginseng induced M2 macrophage to M1 macrophage, increased the accumulation of reactive oxygen species (ROS) and suppressed melanoma tumor growth [56].

Besides natural compounds, lots of small molecules have been proved to regulate macrophage polarization. Metformin was a clinical drug for type-2 diabetes and had been approved to reduce cancer incidence [57], it altered the M2 macrophage and repolarized them into M1 to inhibit tumor growth and metastasis by boosting anti-PD-1 therapy [58], this attributed to the increase of CD8⁺ T cells recruitment into tumors and the decrease of immunosuppressive infiltration of regulatory T cells and other myeloid-derived suppressor cells [58]. Etomoxir was a small molecule and a carnitine palmitoyl transferase-1 (CPT-1) inhibitor, and CPT-1 was a mitochondrial enzyme facilitating the carnitine uptake into the mitochondria matrix for oxidation. High concentration Etomoxir inhibited macrophage polarization by depleting intracellular free coenzyme A (CoA). Pirfenidone was an anti-fiberotic drug used in lung fibrosis treatment, it regulated macrophage polarization and ameliorates the fibrosis, reducing M2 macrophage infiltration and inhibiting TGFβ1-Smad 3 signaling activation [60]. Vitamin D inhibited the macrophage M1 polarization by the STAT-1/TREM-1 pathway [61]. The DNA methyltransferase inhibitor Decitabine in low lose promoted M2 polarization due to the DNA demethylation of PPARγ, and subsequently potentiated the KLF4 binding affinity in immune thrombocytopenia [62]. An M2 polarization inhibitors screening had identified many small molecules selectively inhibited M2 polarization rather than M1 polarization, these molecules including mitogen-activated protein kinase kinase MEK inhibitors trametinib and HDAC inhibitor panobinostat [63].

Moreover, the ultrasound-driven discharge of piezoelectric materials promotes M1 polarization of macrophages [64], and the potentiation of M1 and suppression of M2 polarization was attributed to Ca²⁺ influx through voltage-gated channels and the Ca²⁺-CAMK2A-NF-κB axis, which enhanced the release of the proinflammatory factors TNF-α and IL-1β [64]. But they were still not yet applied in any clinical trials for the anti-tumor functions due to the unknown safety and efficacy.

Oncogenes and epigenetics regulate macrophage polarization

The oncogenes and tumor driving molecules promoted macrophage M2 polarization [65], [66]. For example, the activation of aryl hydrocarbon receptor (AhR) suppressed M1 macrophage polarization and promoted M2 macrophage polarization via the AhR-miR-142a-IRF1/HIF-1α pathway [67]. The high-mobility gene group A2 (HMGA2), an oncoprotein, also promoted M2 polarization and macrophage recruitment by activating STAT3 and inducing CCL2 secretion [68]. Moreover, the transcriptional regulator Krüppel-like factor 4 (KLF4) also regulated macrophage polarization, it was highly expressed in M2 macrophages while low expressed in M1 macrophages, it interacted with STAT6 to induce M2 polarization and inhibited M1 polarization [69]. Conversely, some other proteins promoted macrophage M1 polarization, for instance, the myocyte enhancer factor 2 C (MEF2C) was critical in M1 macrophage polarization regulation in response to infection and inflammation. It promoted M1 polarization and the expression of interleukin-12 p35 subunit (Il12a) and interleukin-12 p40 subunit (Il12b) thus created a pro-inflammatory environment [70]. Moreover, Progranulin (PGRN, encoded by the GRN gene) was a lysosomal and secreted protein [71], it had potential anti-inflammation functions. It inhibited lipopolysaccharide (LPS)-induced macrophage M1 polarization by down regulating NF-кB pathways [72]. Annexin A5 was a soluble protein binding to phospholipid of cell membrane, its high expression promoted macrophage polarization from M1 to M2 by directly interacting with pyruvate kinase M2 (PKM2) at Asp101, Leu104 and Arg106, they formed a tetramer in cytoplasm and blocked PKM2 translocation into nucleus and interaction with HIFα, thus promoted oxidative phosphorylation in an anti-inflammatory status [73]. Similarly, the Annexin A1 also reprogrammed the macrophage towards anti-inflammatory M2 type by AMPK-mTOR pathway [74]. Additionally, the one-carbon metabolism enzyme methylenetetrahydrofolate dehydrogenase 2 (MTHFD2) suppressed IFN-γ and enhanced IL-4 and promoted M2 polarization by targeting PTEN catalytic center (aa 118–141) [75]. STAT6 usually promoted macrophage M2 polarization, but the acetylation of STAT6 by the acetyltransferase CREB-binding protein (CBP) suppressed macrophage M2 polarization [76].

By the way, the protein and RNA modifications regulated macrophage polarization. For example, histone modification regulates macrophage polarization, for example, the histone lysine lactylation induced the M2-like genes in the late phase of M1 macrophages [77], [78]. N6-methyladenosine (m⁶A) was one the most common RNA modifications, it also regulated macrophage polarization [79]. METTL3 as an RNA methyltransferase potentiated the M1 macrophage polarization [80], [81], deficiency of METTL3 suppressed TLR signaling-mediated macrophage activation and increased the M2 macrophages in tumor microenvironment.

Noncoding RNAs regulate the macrophages polarization in cancers

High expression of miR-16 inhibited TAM infiltration and promoted macrophages to M1 polarization [82], while the expression of miR-19a-3p induced macrophages to M2 phenotype and promoted the breast cancer metastasis [83]. Moreover, high expression of long non-coding RNA (lncRNA) ANCR in gastric cancer reduced the level of IL-1β and IL-6 to inhibited macrophage M1 polarization, thus enhanced tumor migration and invasion [84]. In breast cancer, the ectopic expression of lncRNA BCRT1 potentiated the M2 macrophage and tumor metastasis by exosome mediated migration [85]. The expression of lncRNA-COX2 regulated a cascade of immune signals either activate or suppress tumor progression [86], its high expression enhanced the effect of LPS on inflammation and macrophage polarization to M1 phenotype, LncRNA-COX2 restrained sepsis progression by regulating macrophage polarization and inflammatory response by suppressing CREB-C/EBPβ pathway [87]. LncRNA-MM2P was increased during macrophage M2 polarization and downregulated during M1 polarization, its deficiency blocked M2 polarization and thus inhibited the angiogenesis by suppressing STAT6 phosphorylation [88].

Although various methods on polarization modulation have been explored, none of them are in clinical trials now. Current study mainly leverage mice models on the influence of modulating above proteins or metabolites involved in macrophage polarization, thus there is no data on human and all are in pre-clinical stage now.

Progression of CAR-macrophage, advances, challenges and strategies

Even though CAR-T therapy achieved success in hematological malignancies, it confronts difficulties in solid tumors including less T cell infiltration and low trafficking, hard survival for T cells in hostile tumor microenvironment, heterogeneity of tumors, systemic toxicity, and the uncertainty of cancer-specific targets [89], [90]. These inevitable challenges shift our gaze to the engineering of innate immune cells including CAR-NK and CAR-M [91]. NK cells mediate their activity without requiring human leukocyte antigen (HLA) matching, the CAR-NK cells and CAR- NK cells derived from human iPSCs enhanced anti-tumor opportunity in both hematological malignancies and refractory solid tumors [92], [93]. Compared with CAR-T and CAR-NK,CAR-M cells are more easily transported to various tissues through blood vessels and effectively infiltrate solid tumors. CAR-M cells mainly mediate tumor cell phagocytosis in a CAR-dependent and CAR-independent manner. In addition, CAR-M cells also have antigen presentation, co-stimulatory signal transduction, and cytokine secretion abilities, which recruit other immune cells to participate in anti-tumor responses. CAR-M are increasingly be treated as a promising therapeutic method in cancer immunotherapy.

Design and construction of CAR-M

Similar to CAR-T and CAR-NK, CAR-M contains intracellular activation signaling regions, transmembrane regions and extracellular signaling domains to recognize tumor antigens.

The protype of CAR-M can be traced back to Chimeric Antigen Receptors for Phagocytosis (CAR-Ps) engineered in 2018 by Ronald D Vale team [94].They found that the cytosolic domains from Megf10 and FcRɣ robustly triggered engulfment independently of their native extracellular domain, they further connected PI3K subunit and CAR-P- FcRɣ to generate a CAR-P tandem, and found the tandem exhibited good phagocytosis function. A recent study compared these three different CARs and found CAR-M^FcRγ exerted more potent phagocytic and tumor-killing capacity than CAR-MMegf [7]and CAR-M^PI3K98(Fig. 3).

Fig. 3.

Design and development of CAR-M.

Macrophages can be modified with specific CARs to augment antigen presentation and phagocytic ability. The intracellular domain CD3ζ of CAR-M consists of the immunoreceptor tyrosine-based activating motifs (ITAMs). The tandem SH2 in kinase Syk in macrophage binds to CD3ζ and triggers the phagocytosis of macrophage. The CD3ζ-based anti-HER2 CAR macrophages were generated in 2020 [95]. Adenovirus-delivered CAR to macrophage can induce the macrophage to M1 polarization from M2, and induce inflammatory TME and augment the anti-tumor cytotoxicity of T cells, this CAR-M reduced lung metastasis of ovarian cancer [95]. Moreover, the addition of a tandem PI3K recruitment domain in CAR-M augments the phagocytosis [94].

In 2020, induced pluripotent stem cells (iPSCs)-derived CAR-Macrophage (CAR-iMac) has been demonstrated to reduce ovarian cancer tumor size in animal modal [96]. The iPSC cloning was induced with integrable vector encoding reprogramming factors, then the CAR containing CD86 and FcRγ was introduced into the iPSC-introduced macrophage. These CAR-iMacs exhibited M2 phenotype to phagocytize cancer cells without antigen stimulation, while when an antigen such as CD19 is available, the CAR-iMacs will be polarized towards to M1 phenotype. In 2023, a second-generation of iMACs was developed, iMACs with toll-like receptor 4intracellular toll/IL-1R (TIR) domain-containing CARs exhibited better phagocytosis ability than first generation [97]. The novel design of a tandem CD3ζ-TIR dual signaling CAR enhanced engulfment capacity and the kept the macrophage in M1 polarization and resistant to M2 polarization in a nuclear factor kappa B (NF-κB)-dependent manner [97] (Fig. 2).

CAR-147 targeting HER2 is the conjugation of CAR to CD147 transmembrane region and intracellular region. Different from other designs, it doesn’t transmit phagocytic signals, but upregulates the expression of matrix metalloproteinases and thus reduces the collagen level in tumors, to increase T cell infiltration and inhibit breast tumor growth [98]. One obstacle to T cell infiltration is the physical barrier generated by the extracellular matrix (ECM) of solid tumors, while CAR-147 macrophages disrupted the ECM of tumor cells and hence improved the T cell infiltration.

Schematic structures show the the structure of CAR-P, CAR-M, iMAC. All the CARs comprise an extracellular scFV domain to recognize a specific epitope, a transmembrane domain and a hinge region from CD8α. The intracellular signaling domains from CD3ζ, Megf10, PI3K can activate engulfment and they were treated as the 1st generation of CAR-M. The CD3ζ-TIR-CAR containing CD3ζ and TIR in the intracellular domain was defined as the second-generation CAR and possesses higher proinflammatory activity than the first-generation CAR upon binding to a specific antigen.

The advantages and clinical trials of CAR-M

Although CAR-Ms share many common characteristics and obstacles with CAR-T cells, such as the demand for specific antigens, antigen escape and downregulation, and systemic cytokine toxicity(Fig. 4), macrophage engineering has its specific advantages:

• 1)Macrophages are the most abundant immune cells, and the induction of iPSC into macrophages provides an unlimited source of engineered CAR-Ms [96].
• 2)Macrophages have lower toxicity while higher tumor infiltration, thus, they induce T cell infiltration and systemic antitumor immunity to prevent tumor metastasis and prolong survival.
• 3)Macrophages in M2 phenotype are the primary immunosuppressive cells in TME even they suppress other immune cells’ function, they possess the higher phagocytotic ability.
• 4)CAR-M creates a pro-inflammatory environment in tumors and reverses the immunosuppressive TME. Besides the function of phagocytizing tumor cells, it also contributes in antigen presentation and T cell killing ability enhancement.
• 5)CAR-M has shown efficacy in clearing cancer stem cells [99].

Fig. 4.

The crosstalk among macrophage, T cells and dendric cells.

Most of the CAR-Ms are in the pre-clinical stage, but clinical trials have been started too. The CAR-M carrying adenoviral vector Ad5f35 targeting HER2 in different HER2-overexpressing solid tumors has been applied in interventional clinical trials and in Phase I now (NCT04660929, estimation completion time: 2024-12). This multi-center first-in-human trial with CAR-M has shown promise in targeting solid tumors. Another observational cohort study (NCT05007379, estimation completion time: 2023-09) to determine the CAR-macrophages in breast cancer patients’ derived organoids was in progression too, the results haven’t been posted.CT-0508 was the first CAR-M therapy approved by FDA in 2020, and the clinical trial sponsored by Carisma Therapeutics was started in 2021 to treat HER2 positive tumors. However, in March 2024, Carisma Therapeutics company ceased development of CT-0508, a HER2-targeting CAR-Macrophage, given its prioritization of CT-0525 as its lead HER2 product candidate (https://carismatx.com/).

Macrophage are plastic innate immune cells easily to be polarized under different stimulus. M1 macrophages have a better phagocytosis ability and create pro-inflammatory tumor microenvironment. The polarization from M2 to M1 enhanced macrophage phagocytosis. The CAR-M was generated based on different activator in the macrophage to potentiate its anti-tumor capacity. Then new generation of CAR-M enhanced antigen presentation and phagocytosis, and meanwhile created a pro-inflammatory environment which contributes to tumor cells elimination.

Challenges and perspectives of targeting CAR-M therapy

Although CAR-M exhibits promise, it has limitations and challenges: 1) the proliferation ability of macrophages is reduced after injection, leading to the limited amount of macrophages in patients’ blood and affecting the therapy efficacy; 2) the migration ability of the exogenous macrophages is reduced and most of them stay in the liver [100]; 3) antigens may not be enough to be targeted due to the heterogeneous tumor microenvironment. 4) CAR-M is often designed via virus transfection, which may induce insertion mutations. To best utilize CAR-M, specific CAR-M should be designed to maximumly increase the phagocytic ability of macrophages. In addition, the combination therapy of CAR-M and other immunotherapy such as targeting phagocytosis checkpoints exerts better efficacy. Due to the specific advantages of CAR-M and its effective anti-tumor ability in animal experiments, CAR-M as a new cell immunotherapy will be developed fast in the near future, and its efficacy in patients will be shown soon. CAR-M opens another door for cancer immunotherapy in the respect of cell immunotherapy.

Progression of targeting phagocytosis checkpoints

Phagocytosis checkpoints refer to a type of immune checkpoint which are expressed on phagocytes, such as monocyte, DCs and macrophages [101]. They suppress innate immune sensing and work in a similar mechanism to adaptive immune checkpoints such as CTLA-4 and PD-1. Tumor cells usually express the ligand of phagocytosis checkpoints highly to evade immune surveillance, the ligand on tumor cells present a “ don’t eat me” signal and interacts with receptors on the phagocyte, thus the tumor cells achieved “immune escape” successfully. Targeting the immune checkpoint and its ligands or blocking the interaction between phagocytosis checkpoints has emerged as a novel immunotherapy strategy, and therapies on macrophage-mediated phagocytosis checkpoints blockade are the most advanced in clinical trials [15], [102]. We will have a brief summary of all the phagocytosis checkpoints recently reported in cancer immunotherapy, including: CD47-SIRPα, PD-1/PD-L1, CD24-Siglec-10, MHCI-LILRB1, CD22, STC1, and GD2 (Fig. 1, Fig. 5).

Fig. 5.

Phagocytosis checkpoints bridges innate and adaptive immune responses in the tumor microenvironment.

The CD47-SIRPα axis

CD47 is highly expressed on most of the cancer cells, it triggers a “don’t eat me” signal and contributes to the cancer immune evasion by binding SIRPα on macrophages [103](Fig. 5A). The general molecular wiring of CD47-SIRPα in suppressing phagocytosis has been parsed out [104]. CD47 also contributes to drug resistance in lung cancer and positively correlates with cancer cell stemness in glioblastoma. Inhibition of CD47 or blocking its binding with SIRPα potentiates phagocytosis by TAM and reduces the tumor size in a variety of tumors. Clinical trials for drugs targeting CD47 (summarized in Supplementary Table 1) are the most advanced among all the phagocytosis checkpoints. The first-in-human and first-in-class clinical trial of CD47 antibody Hu5F9 in cancers was conducted in 2019 in advanced cancers [105].

Targeting CD47 faces a series of challenges such as anemia, T cell toxicity, unstable blood pressure and transformational difficulty [15], [106]. Strategies are suggested: 1) Designing antibodies based on the differences between cancer cells and red blood cells (RBCs). 2) Targeting SIRPα which doesn’t express on RBCs. ALX148, TTI-621 and ADU-185 were designed based on this concept [107], [108]. 3) Targeting QPCTL, an enzyme catalyzing the pyroglutamylation of CD47, it doesn’t express on RBCs, but is critical for the binding of CD47 and SIRPα [106], [109], [110], [111]. 4) Combination therapies including targeting two or more immune checkpoints or combining with radiotherapy [112], [113]. The combination of Hu5F9-G4 and rituximab enhances macrophage-mediated phagocytosis in B-cell non-Hodgkin’s lymphoma [114]. Chemotherapy or radiotherapy of glioblastoma potentiates CD47 antibody-mediated phagocytosis [115]. 5) Explore other new combination therapies: the integrin activation synergizes the effect of CD47 blockade in cancer immunotherapy [116].

Targeting phagocytosis checkpoints highly rely on the function of tumor-associated macrophages which is a key component of the tumor microenvironment. (A)The binding of CD47 and SIRPα inhibits phagocytosis. RBCs highly express CD47 to avoid phagocytosis by macrophages, targeting CD47 by antibody brings side effects such as anemia or blood palate aggregation as RBCs will be engulfed by macrophages. The high expression of CD47 on immature DCs inhibits its maturation and activation, the high expression of CD47 on T cells leads to T cell apoptosis while on B cells leads to B cell maturation. (B) Targeting PD-1-PD-L1 axis increased the M1 phenotype macrophages and reduced the M2 macrophages. The PD-1-PD-L1 axis facilitates cancer cell immune evasion by not only suppressing T cell function, but also inhibiting phagocytosis by macrophages. The high expression of PD-1 on TAM is associated with the lower phagocytic potency of cancer cells, blocking PD-1-PD-L1 axis augments the phagocytosis and prolongs the murine survival, indicating PD-1-PD-L1 is a novel phagocytosis checkpoint. (C) The high expression of MHC-I on tumor cells binds to LILRB1 on macrophages to inhibit phagocytosis and suppress the antigen presentation from macrophages to T cells. (D) The high expression of CD24 on tumor cells binds to Siglec-10 on T cells to inhibit T cell activation, and binds to Siglec-10 on NK cells to suppress NK cell killing ability; CD24 binding to Siglec-10 on macrophages inhibits phagocytosis. (E) The high expression of STC1 in cancer cell inhibits phagocytosis of cancer cell by macrophage and meanwhile suppress the antigen presentation from APCs to T cells. Thus all the phagocytosis checkpoints are closely associated with the immune cell function and bridge the innate and adaptive immune responses. (F) The high expression of CD22 on B cells in B cell malignancies binds to α 2,6 linked sialic acid on microglia and inhibits phagocytosis of cancer cells by microglia.

The PD-1-PD-L1 axis

PD-L1 is expressed on the surface of tumor cells and PD-1 is expressed on T cells and other immune cells, their binding in tumor microenvironment leads to tumor immune escape and tumor progression. The role of the PD-1-PD-L1 axis in the adaptive immune system has been well explored in previous research [117]. The PD-1-PD-L1 axis facilitates cancer cell immune evasion by not only suppressing T cell function, but also inhibiting phagocytosis by macrophages. TAMs with M2 phenotype express PD-1, they regulate PD-1-PD-L1 immunosuppression by inhibiting T cell activation and function, cytokine secretion and immune exosomes release [118]. Targeting PD-1-PD-L1by monoclonal antibodies have been developed and applied in the clinicals, including both haematological and solid tumors. After immunotherapy of PD-1-PD-L1, the quantity of macrophages were increased too [119], [120],the M1 phenotype macrophages were increased, demonstrated by the increased expression of CD40, MHC-II, iNOS etc. and the M2 phenotype biomarkers were reduced. Meanwhile, the anti-PD-L1 therapy activates macrophages by potentiating the expression of the co-stimulatory molecules CD86 and MHC II [121].

The high expression of PD-1 on TAM is associated with the lower phagocytic potency of cancer cells, blocking PD-1-PD-L1 axis augments the phagocytosis and prolongs the murine survival, indicating PD-1-PD-L1 is a novel phagocytosis checkpoint(Fig. 5B). However, the accurate molecular working mechanism of PD-1-PD-L1 deserves further investigation. Even though drugs targeting the PD-1-PD-L1 axis are already in clinical use, all the drugs were developed based on its function in T-cell immunity instead of phagocytosis in the innate immune response.

The MHC-I- LILRB1 axis

MHC is expressed on all cells except RBCs, primarily functioning in distinguishing “self” and “non-self” in the immune system. The critical function of MHC-I in cancer immunotherapy includes immune response activation and antigen presentation [122], [123], [124], [125], [126]. Therapeutically increased MHC-I expression increased the immune checkpoint blockade efficacy [122]. The expression of MHC-I is regulated by many other molecules. Surface protein sushi domain containing 6 (SUSD6) is a novel negative regulator of MHC-I, it forms a trimolecular complex with TMEM127 and MHC-I, which recruits WWP2 for MHC-I ubiquitination and lysosomal degradation, SUSD6 deletion enhanced MHC-I antigen presentation and reduced tumor growth in a CD8⁺T cell-dependent manner [127]. MHC-I was expressed very few on the Pancreatic cancer cell surface but it was highly accumulated in autophagosomes and lysosomes, thus, inhibition of autophagy restored MHC-I surface expression and enhanced its antigen presentation and following T cell responses [128]. Besides, an evolutionarily conserved function of polycomb repressive complex 2 (PRC2) mediated the MHC-I translational silencing by activating H3K4me3 and repressive H3K27me3 histone modifications, thus promoting cancer cell immune evasion [123]. The MHC-I activation and antigen presentation is also restricted by the activated SUMOylation [129]. Deficiency of histone methyltransferase WHSC1 dampens MHC-I antigen presentation by downregulating MHC-I transcription and suppressing IFN-γ signaling [130].

MHC-I on the surface of cancer cells interacts with the leukocyte immunoglobulin-like receptor subfamily (LILRB) on the macrophages or NK cells to inhibit phagocytosis and helps cancer cell immune evasion [10]. LILRB1 was expressed in most monocytes while LILRB2 expression is much lower. Hence, the MHC-I-LILRB1 axis is another novel phagocytosis checkpoint in cancer immunotherapy [10](Fig. 5C). It is not clear not whether LILRB2 works as phagocytosis checkpoint [102], [131]. Human leukocyte antigen G (HLA-G) is the subclass of MHC-I, it is the ligand of the LILRB1 receptor with the highest affinity. It is highly expressed in cancer cells but rarely expressed in healthy cells. Antibodies targeting HLA-G increase the phagocytosis of tumor cells [132]. Targeting MHC-I blocks the antigen presentation to T cells and T-cell recognition of tumor antigens, thus contributing to tumor immune evasion. Therefore, targeting the MHC-I-LILRB1 axis mainly rely on blocking LILRB1 [133]. More importantly, targeting LILRB1 doesn’t affect hematopoiesis. All the clinical trials targeting the LILRB family are listed in Supplementary Table 1.

CD24-Siglec 10 axis

CD24 is a highly glycosylated membrane protein which was originally termed as the marker for immune cells’ maturation and differentiation. It not only contributes to cancer cell proliferation and metastasis but also facilitates tumor immune evasion. It also induces drug resistance in various cancers.

CD24 has been identified as a phagocytosis checkpoint in 2019 [11] (Fig. 5D). Tumor cells highly express CD24 to avoid innate immune surveillance and augment drug resistance [134]. CD24 on tumor cells binds to the inhibitory receptor sialic-acid-binding Ig-like lectin 10 (Siglec-10) expressed on monocytes, B cells, and activated T cells. Inhibition of the binding of CD24 and Siglec-10 or targeting CD24 by antibodies augments the phagocytosis of cancer cells by TAMs [11], [135]. The Siglec-10 expressed on NK cells inhibited NK cell killing function and promoted tumor cell immune escape, the CD24-Siglec-10 axis between tumor cell and NK cell inhibited cytotoxic ability of NK cells [136]. Siglec-10 on T cells inhibited T cell activation by inhibiting the phosphorylation of T cell receptor related kinases [137], the CD24-Siglec 10 binding between tumor cell and T cell blocked TCR activation [138].

Targeting CD24 has been successful in pre-clinical trials: ALB9 inhibits lung cancer metastasis by targeting CD24 and SWA11 (the monoclonal antibody of CD24) exhibits efficacy in reducing tumor size in many cancers [139]. The monoclonal antibody SN3 and ML5 targeting CD24 induced macrophage phagocytosis in mantle cell lymphoma [140]. The combination use of ML5 and another antibody G7mAb exhibited effective result in HCC [141]. A recent developed nanosphere-based lysosome-targeting chimeras (LYTACs) degradation system integrated with a CD24 antibody inhibited CD24 in tumor and restored macrophage phagocytic function by interrupting CD24-Siglec-10 axis [142]. Furthermore, the combination of nanosphere-anti-CD24 and the glucose oxidase restores macrophage function effectively and inhibits tumor growth without toxicity to normal tissues [143]. However, there are no clinical trials for inhibitors or antibodies targeting CD24 by enhancing the phagocytosis of cancer cells.

CD22

CD22, known as Siglec-2, is exclusively expressed on B cells, it is a cell surface sialoglycoprotein and primarily functions in the regulations of B-cell proliferation and migration. Cytoplasmic CD22 and CD19 are expressed together in the earliest stage of B cell differentiation and appear before CD20 expression. Most precursor B cells also show positive cytoplasmic CD22. In mature B lymphocytes, the expression of surface CD22 precedes or accompanies the expression of surface IgM and/or IgD, but is lost in plasma cells. CD22 is typically a regulatory molecule that negatively regulates B cell receptor signaling and plays a crucial role in establishing B cell activation thresholds. CD22 is highly expressed on the surface of most B-cell malignant tumor cells, including Acute Lymphocytic Leukemia, Non-Hodgkin's Lymphoma, and Chronic Lymphocytic Leukemia. Hence, CD22 is a promising therapeutic targeting of B cell cancers [144].

CD22 on B cells binds to α2,6-linked sialic acid on microglia to suppress the phagocytosis of microglia (Fig. 5F). Blocking CD22 restores the phagocytosis capacity of microglia and enhanced the alternative activation markers such as CD206 and IL-10 expression [12], leading to a reduction in brain lesion and hematoma volume. Thus, CD22 has been recognized as a key molecular switch to control the detrimental effects of microglia after acute brain injury [145]. Drugs targeting CD22 mainly include monoclonal antibodies, an antibody-drug conjugate (ADC) and CAR-T and CAR-NK therapy [146], [147], [148], [149], [150], [151], [152]. The clinical Phase 1b results showed that the closed-loop manufactoring of CD22-CAR T cells is safty and feasible in pediatric and adult B-ALL [153]. Targeting CD22 is widely used to facilitate CD19 or CD20 in CAR-T therapy as it overcomes the antigen loss brought by single-antigen CAR-T therapy [154]. The CD19 and CD22 dual-targeted CAR-T cell therapy (CTA101) for B-ALL has finished Phase I clinical trial and exhibited manageable safety and anti-leukemia effect [155], [156], Phase II clinical trials is under preparation [157], [158], [159]. In the childhood refractory or relapsed B-ALL, the sequential targeting of CD19 and CD22 strategy induced deep and sustained responses with an acceptable toxicity rate(NCT04340154) [160]. Dual targeting CD19 and CD22 with cotransducted T cells prevented antigen-negative relapse after CAR T-cell therapy [161]. The recent report proved that B-ALL patients treated by the allogeneic CD19/CD22 bispecific targeting CAR-T therapy exhibited durable remissions [162].

Moreover, CD22-targeted CD28 bispecific antibody had been approved to enhance the antitumor efficacy of odronextamab in refractory diffuse large B cell lymphoma models [163]. And a novel CD19/CD22/CD3 tri-specific antibody has been demonstrated to overcome immune evasion and potentiate the therapeutic efficacy against B-ALL [164].

The limitation of targeting CD22 maybe the specific efficacy in B-cell lymphomas instead of other tumors as it is exclusively expressed on B cells.

STC1

STC1 is a glycoprotein hormone widely expressed in many tissues in human bodies. It plays a critical role in cancer development. For example, the STC1-Nortch1 signaling promoted the stemness of hepatocellular carcinoma (HCC); STC1 promoted ovarian cancer metastasis by FOXC2/ITGB6 signaling axis [165], [166]; STC1 expressed on the cancer-associated fibroblast promoted colon cancer progression and metastasis by platelet-derived growth factor signaling [167]; STC1 also promoted breast cancer lung metastasis by enhancing EGFA-ERK-S100A4 pathway [168]. Recent study indicated that STC1 enhanced melanoma progression by binding βPIX and this process is mediated by YAP nuclear translocation and M2 macrophage recruitment. STC1 had been treated as a prognostic biomarker for poor prognosis for various cancers [169].

STC1 is an unexpected phagocytosis checkpoint [13]. The interaction between STC1 and calreticulin(an “eat me” signal) inhibits the membrane calreticulin-directed phagocytosis by phagocytes and suppresses the following antigen presentation from APCs to T cells (Fig. 5E). The highly expressed STC1 in tumor cells keeps calreticulin in mitochondria and endoplasmic reticulum (ER), which suppresses the function of macrophage and facilitates cancer immune escape [13]. The phagocytosis inhibition by STC1 leads to cancer immunotherapy resistance. Therefore, inhibition of STC1 or blocking its interaction with calreticulin would be a potential approach to improve cancer immunotherapy efficacy. The inflammation-related molecule A20 reduced the infiltration of immune cells and decreased the “eat-me” signal calreticulin (CRT) protein on cell membrane translocation by enhancing STC1, binding CRT and detaining in mitochondria [170].

GD2-Siglec 7 pathway

GD2 is a disialoganglioside rather than a protein involved in phagocytosis. The level of GD2 is low and mostly enriched in the brain and skin melanocytes, but it is highly expressed in brain tumors. GD2 as a sialoglycan binds to Siglec-7 on macrophages and NK cells specifically. Anti-GD2 or blocking the GD2-Siglec-7 pathway promotes phagocytosis of tumor cells by macrophage by upregulating the expression of calreticulin. Targeting GD2 prolongs the neuroblastoma patients’ survival [171], [172]. Targeting GD2 and CD47 exerts a synergistic effect on neuroblastoma cancer elimination by augmenting the phagocytosis of cancer cells by microglia [14]. Anti-GD2 in other cancers are under pre-clinical research [173], [174], [175], [176]. GD2 is the first ganglioside proved to be a tumor antigen in cancer immunotherapy. Clinical trials of drugs targeting GD2 are in progress [177], [178] (Fig. 6). The GD2-CAR-T therapy targeting H3K27M-mutated diffuse intrinsic pontine glioma is a first-in-human phase I clinical trial [174]. However, targeting GD2 in other tumors besides neuroblastoma is restricted and needs further verifications due to its specificity in neuroblastoma.

Fig. 6.

The potential applications of targeting phagocytosis checkpoints in cancers.

The efficacy of targeting phagocytosis checkpoints is affected by many factors. TME contributes to the failure of cancer immunotherapy [179]. TAMs promote T cell dysfunction and exhaustion to facilitate cancer cell evasion from immune surveillance [180], the high expression of CD47 co-works with the T-cell exhaustion markers PD-1 and CTLA-4 and induces T-cell exhausting and remodeling of the TME. In addition, the expression of the phagocytosis checkpoint may affect each other. i.e. the overexpression of PD-L1 downregulates CD24 in hepatocellular cancer [139].

The fastest clinical trials on macrophage phagocytosis checkpoints are on CD47-SIRPα, there are more than 50 clinical trials on targeting this pathway. However, there are no drugs specifically targeting this pathway used in the clinicals now. Targeting PD-1-PD-L1 has been used in clinical widely due to its function on T cell immune response. Targeting GD2 has been used in neuroblastoma treatment and its combinational therapy with CAR-T is in clinical trial. Other phagocytosis inhibitors specially targeting the phagocytosis pathway are still under development.

Concluding remarks and future perspectives

Targeting macrophages including but not limited to macrophage polarization alteration, CAR-M and targeting phagocytosis checkpoints (Fig. 7). Many other methods are under development too. Macrophage recruitment in tumor tissues is a critical process of immune function, inhibition of the recruitment of circulatory monocytes will be a promising therapy to target TAM [181]. Since CSF-1/CSF-R axis increased the M2 biomarker expression, inhibition of CSF-1 pathway reprogrammed macrophage polarization, the CSF1R inhibitor PLX3397 reprograms macrophages and prolonged tumor growth [182]. Furthermore, M2 macrophages have a better phagocytosis ability in engulfing nanoparticles, therefore, it is an ideal therapy to target M2 macrophages with nanoparticles, for example, the nanoparticles containing TLR agonists (such as TLR7/8) or tumor peptides can promote TAM reprograming and improve the anti-tumor immunity [183]. However, how to specifically deliver these nanoparticles into M2 macrophages and how to continue their long-lasting efficacy inside the macrophages are questions pending to be answered. As the development of nanotechnology, tumor vaccine design improved a lot, tumor vaccine delivery remodeling the TME and potentiate the immunotherapy [184].

Fig. 7.

Main therapies targeting macrophages.

Targeting macrophages is developing fast. Polarization alteration was mostly utilized to change macrophages from M2 to M1 for their anti-tumor function, however, cancer cells also secret colony-stimulating factors which prompt the polarization of macrophages from M1 to M2 phenotype with higher tumorigenicity. How to remodel macrophage from M2 to M1 is the prominent question, how to alter M2 to M1 without affecting other innate immune cells functions also needs to be explored. Furthermore, after the polarization alteration, how to translate it into industry and clinical application is another question. There is still a long way to go regarding to targeting macrophage polarization alteration applications.

The main therapies targeting macrophages include: (1) Inducing macrophages to M1 polarization: Monocytes can be induced to differentiate into M1 or M2 macrophages, and various factors in the tumor microenvironment affect macrophage polarization. The tumor-associated macrophages can be induced into M1 phenotype by different methods. (2) Engineered macrophages: The CAR-Macrophage design and development is already in the clinical trial now, the engineered CAR-M increased its phagocytosis ability. CAR-M release more killing cytokines like INFγ and IL-12 to induce the cancer cell apoptosis and to reprogram the M2 macrophage into M1 macrophage. The nanoparticles not only help the drug delivery therapy but also can be used in targeting macrophage. (3) Blocking the phagocytosis checkpoints: the binding between tumor surface antigens and its receptors in macrophage potentiated tumor immune evasion. Blocking their binding or inhibiting either expression will be promising therapies to target macrophages. (4) Blocking monocyte recruitment and accumulation: M2 macrophage recruitment is a critical process in the tumor microenvironment, cytokines and chemokines contribute to monocyte recruitment, blocking their binding inhibit monocyte recruitment and created a tumor suppressive tumor microenvironment.

This figure was created by Biorender. Website: https://www.biorender.com/.

Compared with CART-T and CAR-NK, CAR-M has its specific advantages as it creates a pro-inflammatory environment and turn the TME into a tumor-suppressive state, it also exerts good anti-tumor efficacy in animal models, but its efficacy, feasibility, tolerability and safety in human are waiting to be tested. The combination of CAR-M and other therapies such as targeting phagocytosis checkpoints will probably augment the function of phagocytes. By combination with PD-1 checkpoint inhibitors, CAR-M could synergize with PD-1 blocker to reprogram the TME and improve overall survival in colon carcinoma-bearing mouse [185]. Blocking phagocytosis checkpoint CD47-SIRPα signal enhances the anti-tumor function of trastuzumab in breast cancer, which directs macrophages to phagocytose cancer cells, suggesting the prospective combination of CAR-M with CD47/SIRPα inhibition [186] (Fig. 6). More CAR-M designs for enhancing phagocytosis function are desired and more antigens need to be explored to apply CAR-M techniques. Following the success of CAR-T in some hematological tumors and failure in some solid tumors, the CAR-M in targeting antigens in solid tumors will overcome the deficiency of CAR-T and exhibit a promising therapeutic effect, especially when combine with other therapies.

Targeting phagocytosis checkpoints bridges innate immunity and adaptive immunity. Inhibition of the phagocytosis checkpoint not only enhances phagocytosis, but also activates the antitumor effect via other pathways such as CD103⁺ DC-NK cell axis in liver cancer when CD47 is blocked [187]. Therefore, targeting phagocytosis checkpoint will be a promising immunotherapy in cancer treatment. But there are also formidable challenges which are pending to be solved: 1) The immunotherapy targeting phagocytosis checkpoints most relies on the phagocytes especially heterogeneous macrophages, the phagocytosis efficacy of cancer cells may vary a lot. 2) The side effects of targeting specific antigens are difficult. e.g. targeting CD47 will lead to anemia. Strategies are developed to overcome this difficulty: targeting the regulation protein of CD47(QPCTL) can avoid this side effect, while the first clinical trials of targeting QPCTL needs to be explored. 3) Inhibitors and antibodies for targeting new phagocytosis checkpoints such as STC1 need to be designed and there will be a long way to go for their clinical application.

Outstanding questions:

• •How to accurately target M2 macrophages without affecting other types of macrophages?
• •How to specifically design CAR-M to target the tumor antigen more accurately?
• •What are the strategies to improve the proliferation and migration of CAR-M after the injection?
• •When will the therapy combing CAR-M with phagocytosis checkpoint targeting enter clinical trial?
• •When the first drug targeting CD47 can be on market and when the first clinical trial targeting QPCTL can be approved in a clinical trial?
• •Will therapy only target CD22 without combination with other therapies effective in B cell malignancies?

Ethics approval and consent to participate

Not applicable.

Consent for publication

The corresponding author has received consent for publication.

Compliance with ethics requirements

This article does not involve any studies with human or animal subjects.

CRediT authorship contribution statement

Yu’e Liu: Writing – original draft. Huabing Tan: Data source, Writing – original draft. Jingyuan Dai: Data source. Jianghua Lin: Data source. Kaijun Zhao: Supervision. Haibo Hu: Conceptualization, Supervision. Chunlong Zhong: Supervision, Conceptualization.

Funding

This review is supported by below grants: the Key Disciplines Group Construction Project of Shanghai Pudong New Area Health Commission (PWZxq2022-10，PW2022A-28), the Medical Discipline Construction Project of Pudong Health Committee of Shanghai (PWYgy2021-07), the Key Discipline Construction Project of Shanghai East Hospital (2024-DFZD- 003) and the Neuroscience, Innovation and Development Research Project (YXJL-2022-00351-0183).

Data Availability

Data and material will be deposited and publicly available. All the clinical trial information is from the website: Https://clinicaltrials.gov. https://www.chinadrugtrials.org.cn/m_index.html.

Declaration of competing interest

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.

Biographies

Dr. Yu’e Liu obtained her doctorate degree from the Grenoble Ecole De Management (France) in 2019, and Ph.D. in basic medicine in Tongji University in 2014. Now she is a post-doc in Harvard Medical School. As a first/co-first author, she has published papers in Advanced Science (2022), Signal Transduction and Targeted Therapy (2020,2023), Journal of Advanced Research (2023), Cellular Oncology (2023), etc. Her research interest is cancer metabolism and cancer immunology.

Pro. Kaijun Zhao, M.D, Postdoctoral Fellow, Associate Chief Physician, and Graduate Supervisor, currently serving as the Deputy Director of the Department of Neurosurgery at Tongji University affiliated East Hospital. He also holds several key positions including the leader of the Vascular Group at Shanghai East Hospital and the Executive Editor for OCIN CASES. Prof. Zhao is an invited reviewer of Stroke and the European Journal of Radiology and published articles in top journals. Dr. Zhao is primarily involved in clinical and foundational research on cerebrovascular diseases, brain tumors, and traumatic brain injuries.

Dr. Haibo Hu got his doctorate degree from Suzhou University in 2017, and now he is a surgeon at the Department of Cardiothoracic Surgery, The Affiliated Huai'an Hospital of Xuzhou Medical University, Huai'an, Jiangsu, China. He has published multiple papers on lung cancer in scientific journals such as J Cell Physiol and Biomed Pharmacotherapy.

Chunlong Zhong is the Chairman of the Department of Neurosurgery at Shanghai East Hospital, School of Medicine, Tongji University. His clinical expertise focuses on the surgical treatment of patients with head trauma, stroke, or brain tumors, with a special interest in the multidisciplinary team treatment of pituitary adenoma and endoscopic skull base surgery. He has led several national fund projects focusing on investigating novel therapeutic strategies for traumatic brain injury (TBI) and neuro-oncology.

Appendix A. Supplementary data

The following are the Supplementary data to this article: