1. Introduction
Malignant tumor is a major health problem worldwide and a leading cause of death, ranking second only to cardiovascular disease. Immunotherapy based on the tumor microenvironment (TME) has become a promising cancer treatment strategy (1). The TME is a complex system composed mainly of tumor cells, mesenchymal cells, innate immune cells [macrophages, monocytes, natural killer (NK) cells, dendritic cells and myeloid derived suppressor cells] and adaptive immune cells (T cells and B cells). Macrophages, also known as tumor-associated macrophages (TAMs), among innate immune cells [macrophages, monocytes, natural killer (NK) cells, dendritic cells and myeloid derived suppressor cells], which are found in almost all tissues and organs, are the first line of defense against exogenous and endogenous injury- or pathogen-associated molecular patterns (2). Macrophages that infiltrate tumor tissue or congregate in the microenvironment of solid tumors are defined as tumor-associated macrophages (TAMs). As an important part of the TME, macrophages have a complex role in tumorigenesis and tumor progression. They can not only inhibit tumor growth by releasing pro-inflammatory cytokines and exerting cytotoxic activities, but also promote tumor progression by affecting the occurrence and growth of tumor cells, participating in tumor angiogenesis and metastasis, and shaping an immunosuppressive microenvironment (3). In the present review, the heterogeneity of the origin of TAMs, the factors affecting the polarization of TAMs and the complex role of TAMs in tumor progression were summarized, and therapeutic approaches targeting TAMs, such as consuming TAMs or re-educating TAMs were discussed in order to provide a reference for colleagues to gain insight for tumor immunotherapy.
2. Origin and classification of TAMs
Origin
It is generally thought that TAMs are mainly derived from monocytes produced by hematopoietic stem cells in bone marrow (4). However, recent evidence suggested that certain TAMs (such as alveolar macrophages, brain macrophages and liver macrophages) originate from pre-natal embryonic precursors (yolk sac or fetal liver) (5). These cells were recruited into the TME under the action of chemokines [such as C-C motif ligand 2 (CCL2), CCL3, CCL4 and CXCL12], colony-stimulating factor 1 (CSF-1), interleukin-6 (IL-6), IL-1β and vascular endothelial growth factor (VEGF) produced by tumor cells or stromal cells to become TAMs (6,7). As an important part of the TME, these TAMs can affect tumorigenesis and development, tumor angiogenesis and immune regulation through interactions with other cell populations in the TME. Furthermore, TAMs are also associated with poor prognosis of various tumors, such as breast cancer (8), bladder cancer (9), head and neck neoplasm (10), glioma (11), melanoma (12) and prostate cancer (13). However, more recently, it has been found that high macrophage infiltration is associated with better prognosis in colorectal and gastric cancers (14). This opposite effect may be related to the plasticity of macrophages and the resulting heterogeneity in phenotype and function of various cancers.
Classification
Macrophages are highly plastic and their phenotype and function are regulated by the surrounding microenvironment. Macrophages usually exist in two distinct subsets: Classically activated macrophages (M1) and alternately activated macrophages (M2). M1 macrophages, normally recognized and induced by type I T-helper cell (Th1) cytokines [e.g. IFN-γ and tumor necrosis factor (TNF)-α] or bacterial lipopolysaccharide (LPS), secrete pro-inflammatory cytokines such as IL-12 and TNF-α, produce high levels of nitric oxide (NO) and reactive oxygen species (ROS), and have powerful anti-microbial and anti-tumor activities (15). By contrast, M2 macrophages are activated by Th2 cytokines (IL-4 and IL-13), secrete anti-inflammatory cytokines such as IL-10, IL-13 and IL-4, and express abundant arginase-1 (Arg-1), mannose receptor (CD206) and scavenger receptor (CD163), which have the functions of removing debris, promoting angiogenesis, tissue reconstruction, damage repair, as well as promoting tumorigenesis and development (16). However, depending on the different activation stimuli, M2 macrophages can be further divided into four distinct subgroups, including M2a, M2b, M2c and M2d. The M2a subgroup is induced by IL-4 and IL-13 to produce high levels of CD206, decoy receptor IL-1 receptor II and IL-1 receptor antagonist (17,18). The M2b subgroup can be induced by immune complexes combined with IL-1β or LPS, which produces anti-inflammatory and pro-inflammatory cytokines IL-10, IL-1β, IL-6 and TNF-α (17,18). The M2c subgroup is generally induced by glucocorticoids and IL-10, releases large amounts of IL-10 and TGF-β and exhibits strong anti-inflammatory activity against apoptotic cells (17,19). Finally, the M2d subgroup, induced by Toll-like receptor (TLR) agonists via adenosine receptors, inhibits the production of pro-inflammatory cytokines and induces the secretion of anti-inflammatory cytokines and VEGFs (20-22). Overall, M2a and M2b macrophages have an immunomodulatory role and promote helper T-cell responses, while M2c macrophages are associated with immune response suppression and tissue remodeling. M2d macrophages are involved in angiogenesis and tumor progression (17,23) (Fig. 1).
3. Polarization of macrophages
The factors influencing TAM polarization are diverse and regulated by a variety of signals in tumor cells and stromal cells in the TME, mainly including the following factors, immunogenic signals, hypoxia (tumor-derived metabolic signals) and extracellular matrix (ECM) components.
Immunogenic signals
Immunogenic signaling mainly refers to the cytokines, chemokines and growth factors released by tumor cells, stromal cells and other infiltrating cells in the microenvironment, which are key determinants of TAM polarization. Among these factors, CCL2 and CSF-1 are the most well-studied stimulators. Studies have shown that tumor-derived chemokine CCL2 binds to C-C chemokine receptor 2 (CCR2) expressed on the surface of macrophages to polarize macrophages towards the pro-tumor phenotype. Blocking the CCL2-CCR2 interaction by gene ablation or antibodies was observed to significantly inhibit tumor metastasis, prolong the survival of tumor-bearing mice and reduce the expression of pro-tumor cytokines (24-26). Later, it was proved that, in addition to the influence of CCL2 derived from tumor cells on the polarization of macrophages, the chemotactic signal of CCL2 expressed by tumor-associated fibroblasts could also recruit macrophages to the tumor site and drive the polarization of macrophages towards the pro-tumor phenotype to increase the aggressiveness of cancer cells and the occurrence of breast tumors (27). CSF-1, another important factor affecting the polarization of macrophages, is generally widely overexpressed at the invasive margins of various tumors and is associated with tumor progression (15). A study on follicular lymphoma found that CSF-1 derived from tumor cells promoted the polarization of macrophages toward pro-tumor M2-like phenotypes and the use of CSF-1R inhibitor (PLX3397) resulted in the repolarization of macrophages toward the M1-type, which showed synergistic anti-tumor effects when combined with anti-CD20 rituximab (28). In addition to the above factors, there are other factors involved in the induction of macrophage polarization, such as VEGF-A, epidermal growth factor (EGF) and prostaglandin E2 (29-31).
Hypoxia
Hypoxia, as one of the key drivers of macrophage recruitment and polarization in the TME, is caused by tumor cells with vigorous metabolism and rapid growth but poor vascular organization, which is a common feature of most solid tumors (32). Hypoxia can regulate the phenotype of TAMs through various factors, particularly through lactic acid produced by glycolysis affecting macrophage polarization (33). For instance, a study on gastric cancer showed that lactic acid produced by glycolysis can induce changes in the macrophage phenotype through monocarboxylate channel transporter-hypoxia-inducible factor 1α signaling, polarizing macrophages towards an M2-like state (34). In addition, the hypoxic TME can stimulate the secretion of tumor-derived exosomes, regulate the macrophage phenotype and promote tumor development. For instance, exosomes of hypoxic tumor cells are enriched with immunomodulatory proteins and chemokines, including CSF-1, CCL2, ferritin heavy chain, ferritin light chain and TGF-β, which influence macrophage recruitment and promote macrophage polarization to the M2 phenotype (35). Therefore, the hypoxic microenvironment can shape a specific macrophage phenotype, which promotes immune escape and metastasis of tumor cells.
ECM components
Last but not least, the ECM component of the TME also has a regulatory effect on macrophage polarization. The ECM is a highly dynamic and complex macromolecular network consisting of a variety of fibrins, proteoglycans and stromal cell-associated proteins (36). The ECM molecule, elastin microfibril interfacer 2 (EMILIN-2), which belongs to the EDEN protein family, exerts a tumor-suppressive effect by affecting the polarization state of macrophages in a variety of tumors (37,38). In a study on colorectal cancer, EMILIN-2 was found to promote M1 polarization through activation of the TLR-4/myeloid differentiation factor-88/NF-κB signaling pathway. EMILIN-2 deficiency is associated with increased M2 macrophage infiltration (39). This suggests that certain components of the ECM are key regulators of the tumor-associated inflammatory environment and may represent promising prognostic biomarkers for tumor patients. In summary, the polarization of TAMs is regulated by complex biological networks, which is closely related to cancer development. Understanding the mechanisms of macrophage polarization may enable researchers to manipulate the macrophage polarization status to stimulate their anti-tumor potential for therapeutic purposes.
5. Therapeutics targeting macrophages
TAMs are abundant in the TME of most cancer types and are commonly associated with poor clinical prognosis for cancer patients. TAMs are becoming a key target for immunotherapy and the different approaches targeting TAMs that have been explored may be broadly divided into three main categories: i) Eliminating TAMs already present in the TME; ii) inhibiting monocyte recruitment; iii) reprogramming of TAMs (67,68). These strategies have been investigated in preclinical models and some of them have been translated into clinical studies as adjuncts to immunotherapy (69). In the present review, some of the current approaches to targeting macrophages and clinical trials were outlined and the potential advantages and disadvantages of these approaches were discussed (Table I).
Consumption of macrophages already present in the TME
Selective elimination of TAMs has been used in cancer treatment. An attractive strategy for depleting TAMs in the TME is to trigger its apoptosis and restore local immune surveillance in the TME, which can effectively inhibit tumor growth. Several compounds have been shown to induce apoptosis in macrophages, mainly including bisphosphonates and trabectedin.
Bisphosphonates are a class of anti-bone resorption drugs, which can be divided into two categories according to their structural characteristics: Non-nitrogenous bisphosphonates and nitrogenous bisphosphonates (70). Bisphosphonates exhibit direct or indirect antitumor properties. They can inhibit cancer cell proliferation, induce tumor cell apoptosis, block angiogenesis and interfere with immune surveillance. At the same time, bisphosphonates can also inhibit the proliferation, migration and invasion of macrophages, leading to the apoptosis of macrophages (71,72). For instance, in earlier studies, dichloromethylenediphosphonates (also known as clodronates) from the non-nitrogenous bisphosphonate family were often used to consume macrophages in the liver and spleen when loaded with liposomes (73); Zoledronate, the third generation of nitrogenous bisphosphonates, is selectively cytotoxic to TAMs expressing MMP9 and inhibits the progression of cervical cancer (74). In addition, a study of non-small cell lung cancer found that calcium zoledronate nanoparticles modified with biotin and mannose preferentially targeted biotin-expressing tumor cells and mannose-expressing TAMs, ultimately suppressing tumor growth and survival (75). A study about thyroid cancer found that zoledronic acid inhibits thyroid cancer stemness and metastasis by repressing M2-like TAM polarization and the Wnt/β-catenin pathway, reducing the tumor burden (76). Furthermore, a prospective phase II clinical trial found that zoledronic acid combined with radiotherapy reduced bone pain and improved quality of life in patients with bone metastases from gastrointestinal tumors (77). Trabectedin is a tetrahydroisoquinoline alkaloid that directly kills tumor cells by interfering with multiple transcription factors, DNA-binding proteins and DNA repair pathways (78). Furthermore, it also selectively consumes monocytes and macrophages in the TME by activating caspase 8 through a TNF-related apoptosis-inducing ligand-dependent mechanism (79). In a study on fibrosarcoma, Trabectedin selectively reduced macrophages in the TME and enhanced the anti-tumor response to anti-PD-1 therapy (80). In several clinical trials, Trabectedin was found to show good safety and efficacy in the treatment of soft tissue sarcoma and ovarian cancer (81,82). However, disappointing results were obtained in malignant pleural mesothelioma and pancreatic cancer tumors (83,84).
Macrophages have an important regulatory role in maintaining host defenses and tissue homeostasis, and a major problem with the depletion of macrophages is the inability to avoid the side effects that arise from non-selective macrophage depletion. Therefore, the key to minimizing potential toxic side effects is to develop drugs that preferentially target M2-like macrophages. A recent study designed M2-targeting nanoliposomes, which effectively depleted M2-type TAMs, remodeled the TME and effectively inhibited tumor growth (85). Consequently, targeted elimination of M2-like TAMs is a promising approach for cancer immunotherapy.
Inhibition of monocyte recruitment
As mentioned earlier, most TAMs originate from the production of bone marrow monocytes. TAMs are recruited to tumor sites in response to tumor-derived chemokines. Therefore, the application of monoclonal antibodies or small molecule inhibitors to interfere with chemokine signaling may be an effective way to prevent the accumulation of TAMs in the TME.
The expansion of TAMs in tumors is usually mediated by monocyte recruitment on the CCL2-CCR2 axis. Monocytes expressing the CCR-2 receptor are recruited to the tumor site by CCL-2 released by tumor cells, macrophages and stromal cells within the TME, where they further mature into TAMs (86). Therefore, reducing macrophage recruitment and infiltration into the TME by blocking the CCL2/CCR2 axis may be a promising therapeutic anti-tumor strategy. A study on esophageal cancer found that blocking the CCL2-CCR2 axis significantly reduced the tumor incidence by preventing TAM recruitment and also enhanced the anti-tumor effects of CD8+ T cells in the TME (87). Several CCL-2 antibodies are undergoing clinical trials. The two main drugs tested so far are the anti-CCL2 monoclonal antibody Carlumab (CNTO 888) and a targeted small molecule inhibitor (PF04136309), which have shown a certain benefit in tumor control (88,89).
In addition, BMS-813160, a CCR2/5 inhibitor, has been selected as a clinical candidate for its ability to inhibit the migration of inflammatory monocytes and macrophages. Clinical trials of BMS-813160 are ongoing in non-small cell carcinoma, liver cancer and pancreatic cancer (90) (Table I).
CSF-1 controls the proliferation, differentiation, recruitment, survival and function of mononuclear phagocytes (e.g., macrophages, monocytes) (91). Therefore, targeting the CSF1/CSF1R signaling pathway is considered to be another important and effective strategy for the treatment of malignant tumors. Ries et al (92) found that the use of CSF1R monoclonal antibody, RG7155, in patients with advanced diffuse giant cell tumors significantly reduced CSF1R+CD163+ macrophages in the TME. In a study of advanced solid tumors, RG7155 specifically depleted immunosuppressed M2-like macrophages and was used in combination with paclitaxel to enhance anti-tumor responses (93). In addition, in a study on endometrial cancer, it was found to promote TAM recruitment in the TME and tumor cell proliferation, which was significantly diminished by a CSF1R blocker (PLX3397) (94). A phase I clinical trial found PLX3397 to have favorable safety and tolerability in Asian patients with advanced solid tumors (95). In a rare tumor called tendonsynovial giant cell tumor, vimseltinib, a small molecule inhibitor targeting CSF1R, was found to persistently inhibit CSF1R activity in vitro and in vivo, depleting macrophages and other CSF1R-dependent cells, and inhibiting tumor growth and bone degradation in a mouse model of cancer (96).
Although CSF1/CSF1R and CCL2/CCR2 blockade are the most widely studied axes of TAM depletion, other cytokines have also been shown to have a role in this process. In a mouse model of malignant lobular tumors, monocytes are recruited into tumors through the interaction between CCL5 and CCR5. Blockade of the CCL5-CCR5 axis by CCR5 inhibitors resulted in markedly attenuated monocyte recruitment into tumors and inhibition of tumor growth (97). Therefore, actively exploring the factors that interfere with macrophage recruitment will provide a new idea for targeted macrophage therapy.
TAM reprogramming
As mentioned above, it is widely acknowledged that M2 and M1 macrophages have opposite roles in tumor growth and metastasis. Therefore, therapeutic strategies to re-educate the tumor-promoting M2 phenotype into the tumor-killing M1 phenotype have been proposed. Reprogramming macrophages involves the following two aspects: Restoration of phagocytosis and promotion of the polarization phenotype.
Role of cellular phenotypes in restoring phagocytosis
i) CD24
CD24 is a highly glycosylated glycosylphosphatidylinositol-anchored surface protein that acts as a 'don't eat me' signal. It regulates phagocytosis in macrophages through interaction with sialic-binding IG-10 (Siglec-10), an inhibitory receptor on TAMs (98). CD24 is commonly overexpressed in cancers and its overexpression is associated with poor prognosis in various cancers (99). It has been found in ovarian cancer and TNBC that CD24 expressed by tumor cells can interact with Siglec-10 on TAMs, blocking phagocytosis of macrophages. Blockade of CD24 or Siglec-10 enhanced the phagocytosis of macrophages on tumors and inhibited tumor growth (100). In addition, a recent study in mantle cell lymphoma and follicular lymphoma found that high expression of CD24 was associated with poor prognosis of patients. Treatment with CD24 monoclonal antibody significantly enhanced phagocytosis by macrophages and inhibited tumor progression (101). These studies demonstrate the therapeutic potential of CD24 blockade as a cancer immunotherapy.
ii) CD47
CD47, a transmembrane glycoprotein widely expressed on cancer cells, can block macrophage phagocytosis by binding to the signal-regulatory protein (SIRP)α on the surface of macrophages, enabling cancer cells to escape immune surveillance (102). Based on this characteristic, targeting the CD47-SIRPα axis is an effective modality. In a study of thyroid cancer, it was found that the degree of infiltration of TAMs in xenografted mice treated with anti-CD47 antibody was significantly increased and phagocytosis of tumor cells by macrophages was enhanced, which inhibited tumor growth (103). Furthermore, in small cell lung cancer (SCLC), macrophages showed increased phagocytosis and enhanced anti-tumor effects on GFP-expressing SCLC cells in mice treated with radiotherapy combined with CD47 block, compared with CD47 block alone, a finding that is particularly important for cancer patients suffering from metastatic disease (104). In addition to increasing phagocytosis of tumor cells, anti-CD47 therapy has also been shown to modulate TAM phenotypic changes. For instance, a glioblastoma study found that anti-CD47 treatment not only enhanced macrophage phagocytosis of tumor cells but also shifted the phenotype of macrophages towards the M1 subtype (105). Hence, preclinical studies of CD47-SIRPα blockade suggest its potential for clinical efficacy. To date, clinical studies have indicated that CD47 inhibitors (Hu5F9-G4) in combination with rituximab showed good anti-tumor activity in patients with non-Hodgkin lymphoma, while CD47 inhibitors (CC90002) have provided disappointing results in patients with acute myeloid leukemia (106,107). Clinical trials targeting SIRP molecules are still ongoing (Table I).
Use of polarization changes for therapy
i) PI3Kγ
As a member of the class 1B family, PI3Kγ is usually associated with G-protein-coupled receptor signaling and is abundantly expressed on a variety of immune cells, including macrophages and neutrophils, and the PI3Kγ pathway is related to the phenotypic transformation of TAMs as well as immunosuppressive states (108,109). For instance, in a mouse model carrying breast cancer, blocking PI3Kγ was also found to reduce the number of M2-like macrophages and enhance the role of cytotoxic T lymphocytes (CTLs), which significantly prevented tumor progression and prolonged survival (110). In addition, Carnevalli et al (111) found that AZD3458, a PI3Kγ inhibitor, did not alter tumor macrophage polarization, but instead promoted antigen-presenting macrophage and cytotoxic macrophage activation, which activated CD8 T cell-mediated antitumor activity of associated immune checkpoint inhibitors. A phase I clinical study found that a highly selective PI3Kγ inhibitor (IPI-549) demonstrated significant anti-tumor activity in patients with advanced solid disease (112). Therefore, the above studies have demonstrated the important role of PI3Kγ in TAMs reprogramming and recruitment of immune cells that inhibit tumor growth, suggesting that PI3Kγ may be a promising target for tumor therapy.
ii) TLRs
TLRs are innate immune pattern recognition receptors that have an important role in the activation of innate immune responses (113). In the reprogramming of macrophages, the phenotype of macrophages can change and they may be polarized in a pro-inflammatory direction. Numerous therapies aim at targeting TLR to repolarize macrophages from an M2-like activated state to an M1-like activated state and to enhance the immune response to cancer cells. For instance, local in situ inoculation of melanoma and neck tumor models with the tumor antigen (protein and peptide) adjuvant nanoemulsion loaded with TLR7/8 agonists induced recruitment and activation of innate immune cells, infiltration of lymphocytes and polarization of tumor-associated M2 macrophages, resulting in inhibition of tumor growth and prolonged tumor survival (114). Furthermore, oxaliplatin combined with TLR agonist R848 was found to reverse the polarization of macrophages and enhance the anti-tumor effects of oxaliplatin in colorectal cancer resistant to oxaliplatin (115). In addition to TLR7 and TLR8, it was found that TLR3 and TLR4 agonists also have the effect of altering the macrophage polarization status (116,117). These findings suggest the clinical relevance of TLR signaling and its potential application for cancer treatment by targeting TAMs. To date, several targeting TLRs have been tested in clinical trials with promising results (118-120).
iii) CD40
CD40 is a member of the TNF receptor superfamily and is highly expressed on antigen-presenting cells, such as macrophages. The ligand of CD40 is CD40L, which is mainly expressed by activated T cells, B cells and platelets (121). The CD40-CD40L interaction promotes the polarization of macrophages towards pro-inflammatory macrophages. For instance, in a mouse model of melanoma, it was found that myeloid-derived suppressor cells induce macrophage reprogramming by inhibiting the expression of CD40 on the surface of macrophages, thereby promoting melanoma progression (122). In addition, a study on pancreatic cancer found that CD40 agonists could drive the transformation of the macrophage phenotype towards M1, remodeling the pancreatic cancer microenvironment and inhibiting the development of pancreatic cancer (123). More recently, Frankish et al (124) found that HERA-CD40L, a novel molecule targeting CD40-mediated signal transduction, activates the signaling transduction mechanisms in dendritic cells, leading to an increase in intratumoral T cells and manipulating TME phenotypic changes to repolarize M2 macrophages to M1, thereby enhancing tumor control. In addition, a clinical trial in pancreatic ductal adenocarcinoma showed that selicrelumab, an agonist CD40 antibody, induced changes in the TME in patients with resectable pancreatic cancer, a reduction in M2-like TAMs and a significant reduction in tumor burden (125). Based on the above observations, reprogramming TAMs to anti-tumor phenotypes using CD40, rather than targeting and ablating TAMs, may be the preferred therapeutic paradigm for cancer treatment.
6. Conclusion and prospects
TAMs are one of the most important innate immune cell types in TME and have a complex regulatory role in tumor therapy. Therefore, it is crucial to reveal the exact regulatory mechanisms and specific targets of macrophages on tumors in order to optimize the effectiveness of current tumor therapies. In the present review, the heterogeneity of the origin of TAMs, the relevant regulators of recruitment and polarization and the complex roles of TAMs in tumorigenesis, progression and metastasis were discussed. Certain therapeutic approaches targeting TAMs, such as consumption of TAMs or re-education of TAMs were also outlined to provide insight for tumor immunotherapy. Specifically, targeting TAMs is a promising immunotherapy strategy. However, the clinical application of current therapeutic strategies is still very limited. For instance, the efficacy is restricted to certain patients, the anti-tumor spectrum is narrow, the adverse reactions are more frequent and drug resistance may easily occur. These defects limit the clinical application of targeting macrophages in tumor therapy. In addition, numerous questions remain regarding the nature and function of macrophages in the TME; many unknown molecular mechanisms have a crucial role in regulating tumor growth and development, and various potential targets require more research and attention. Therefore, it is necessary to further investigate the dialogue between macrophages and tumor cells. With a greater understanding of macrophage diversity through single-cell sequencing and other techniques, TAM-targeted therapies will be an important complement to cancer immunotherapy.