Immunotherapy is one of the most effective approaches for the treatment of cancer. Immune checkpoint blockade (ICB) therapies, such as programmed cell death protein 1 (PD-1)/programmed cell death-ligand 1 (PD-L1) inhibitors and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) inhibitors, have exhibited promising therapeutic effects in the treatment of various malignancies, such as non-small cell lung cancer (NSCLC) and melanoma. Moreover, the field of immunotherapy has recently entered a new era with the emergence of novel immunotherapeutic strategies, such as chimeric antigen receptor T cells (CAR-Ts) and CAR macrophages (CAR-Ms) (1). However, challenges in immunotherapy based on immune checkpoint inhibitors (ICIs) still persist. Clinical trials have demonstrated that the response rates for immune drug monotherapy are low, ranging from 10% to 30% for most solid tumors, and that responses widely vary among tumors or individuals (2–4). Moreover, challenges in tumor immunotherapy, such as severe side effects and a lack of tumor-infiltrating lymphocytes, remain (5). Although adverse events associated with immunotherapy may indicate activation of the immune system, severe toxicity can be fatal (1). Therefore, developing strategies to improve clinical response rates and reduce drug toxicity is crucial.

Mounting evidence suggests that immunosuppression in the tumor microenvironment (TME) is a major challenge in maximizing the clinical efficacy of immunotherapy. The TME comprises a complex population of non-tumor cells that influence tumor immune evasion and affect the response rate and survival time to immunotherapy (6), including cancer-associated fibroblasts (CAFs) (7), endothelial cells (8), and immune cells, such as T cells and macrophages (9), etc. Macrophages exhibit high plasticity, whereby alterations in the environment can lead to alterations in the relative abundance of their two primary phenotypes, namely M1 and M2 macrophages. M1 macrophages, also known as classically activated macrophages, exhibit high expression of CD86, inducible nitric oxide synthase (iNOS), and other markers. Lipopolysaccharide and pro-inflammatory cytokines, such as interferon-γ (IFN-γ), can trigger M1 macrophages to secrete tumor necrosis factor-α (TNF-α), pro-inflammatory cytokines, including interleukin-1 (IL-1), IL-6, IL-8, and IL-12, and promote anti-tumor immune response. M2 macrophages, also known as alternatively activated macrophages, exhibit high expression of CD163, CD206, and other markers. Immunoregulatory cytokines, like IL-4, macrophage colony stimulating factor (M-CSF) can trigger M2 macrophages to secrete immunosuppressive factors, such as arginase 1 (Arg-1), IL-10, and transforming growth factor-β (TGF-β), and mediate the formation of immunosuppressive microenvironment (10–12).

Tumor associated macrophages (TAMs) recognized as one of the most crucial components of the TME, accounting for over 50% of the immune cells that infiltrate the TME (13, 14). The phenotypes of TAMs are extremely complex and diverse and M2 macrophages constitute the predominant TAM subpopulations. Moreover, certain TAM subpopulations may express both M1 and M2 signature molecules. Therefore, researchers mostly differentiate TAMs as M1-like and M2-like macrophages. TAMs have been implicated in a variety of processes in: (1) promotion of tumorigenesis and progression through cancer-associated inflammation (15), (2) introduction of tumor metastasis through involvement in epithelial-mesenchymal transition (EMT) regulation (16), angiogenesis and remodeling, and tumor cell intravasation (17–19), and (3) inhibition of T cell activity, secretion of cytokines such as IL-10, and expression of immune checkpoint ligands, thereby promoting immunosuppression of TME (20, 21). Additionally, TAMs regulate the anti-tumor immune response and are associated with low immunotherapeutic efficacy and decreased overall survival (OS) of patients with tumors (22, 23). Furthermore, the abundance of immune checkpoints in TAMs is closely related to the prognosis of immunotherapy in patients (24, 25). In addition, targeting immune checkpoints in TAMs enhances immunotherapeutic effects (26). These findings contribute to our understanding of immune checkpoints in TAMs and their role in immunotherapy, which is essential for developing effective immunotherapeutic strategies.

In this review, we summarize the functions and regulatory pathways of immune checkpoint expression in TAMs. The immune checkpoint molecules discussed include PD-1, PD-L1, signal regulatory protein α (SIRP-α), CD39, CD73, V-type immunoglobulin domain-containing suppressor of T-cell activation (VISTA), and sialic-acid-binding Ig-like lectin 10 (siglec-10) ( Figure 1 ). This review also presents an overview of strategies aimed at sensitizing immune checkpoint therapy through the application of drugs, inhibitors, and nanoparticles, as well as potential approaches to address the poor efficacy of immunotherapy, which are crucial for developing novel immunotherapeutic strategies.

PD-1, also known as CD279, is a member of the B7/CD28 family and is primarily expressed in activated CD4⁺ T cells, CD8⁺ T cells, B cells, NK cells, and other immune cells (27). PD-L1, also known as CD274 or B7-H1, is a ligand of PD-1 and is widely expressed in most tumor cells, macrophages, activated T cells, B cells, monocytes, and endothelial cells. The binding of PD-1 to PD-L1 recruits SH2 domain-containing protein-tyrosine phosphatase-2 (SHP-2) via the immunoreceptor tyrosine-based switch motif. This event subsequently triggers the dephosphorylation of spleen tyrosine kinase and PI3K, which then results in the inhibition of downstream AKT, ERK, and other signaling pathways; thereby suppressing the function of initial and effector T cells, inducing the production and activity of regulatory T cells, promoting tumor progression, and ultimately leading to a poor prognosis (28, 29).

The Food and Drug Administration (FDA) has approved numerous monoclonal antibodies (mAbs) against PD-1 and PD-L1 (nivolumab, pembrolizumab, atezolizumab, and durvalumab) for the treatment of various cancers, including melanoma and NSCLC (30–32). For example, pembrolizumab is used as the first-line treatment for lung adenocarcinoma (LUAD) or large-cell lung cancer with a PD-L1-positivity of ≥50%, and atezolizumab is used as the first-line treatment for patients with LUAD or sarcoma. Moreover, a combination of pembrolizumab and pemetrexed is used as a first-line treatment for patients exhibiting 1%–49% PD-L1 expression (33). ICI therapy has been reported to be beneficial for patients with PD-L1-positive NSCLC (30, 34). However, some patients with NSCLC strongly expressed PD-L1 and did not respond to PD-1/PD-L1 antibody therapy (35). In addition, patients with certain malignancies, such as breast, colorectal, renal cell, and prostate cancers exhibited inadequate response or low response rates to PD-1/PD-L1 antibody therapy (36–40). The reasons are related to various factors, such as the increasing expression of PD-L1, owing to various factors, such as the overexpression of PD-L1 (41), the lack of effective antigen presentation (42, 43), and the influence of TME, such as that of the Tregs, TAMs, and myeloid-derived suppressor cells (MDSCs). Furthermore, in clinical trials, hyper-progression has been observed, which indicated accelerated disease progression following ICB therapy, leading to poor prognosis (44, 45). The immunological mechanisms underlying hyper-progression are complex and may be primarily due to the destabilization of the immune editing and escape phases of tumors by ICIs, which ultimately promotes immune escape and results in accelerated tumor growth (46). ICB therapy is beneficial for patients with PD-L1-negative tumors, including NSCLC, kidney cancer, and other tumors, potentially owing to higher mutational and neoantigen load, possibly because of the heterogeneous expression of PD-L1 in the TME and distinct tumor histology (47, 48).

Thus, evaluation of tumor PD-L1 in isolation does not provide a comprehensive and accurate prediction of the effect of immunotherapy. Singhal et al. proposed that PD-L1 alone is an imperfect diagnostic biomarker (49). Therefore, combined analysis of various predictors, such as the addition of detecting macrophage PD-L1 levels, is needed to more faithfully reflect the composition of PD-L1 in the TME, predict the efficacy of PD-1/PD-L1 blockade, and maximize the therapeutic effects of immunotherapy in patients with cancer.

CD47, also known as integrin-associated protein, is widely expressed in normal cells. CD47 ligands include signal regulatory protein α (SIRP-α), thrombospondin-1 (TSP-1), and integrins (131). SIRP-α is a member of the SIRP family, which constitutes of transmembrane glycoproteins, and its binding to CD47 is crucial for macrophage phagocytosis regulation. Upon binding of CD47 to SIRP-α in macrophages, two typical immunoreceptor tyrosine-based inhibitory motifs (ITIMs) at the tail end of SIRP-α are phosphorylated. Phosphorylated ITIM recruited and activated tyrosine phosphatases, particularly SHP⁃1 and SHP⁃2, which inhibited myosin accumulation in phagocyte synapses and limited the phagocytic activity of macrophages, thereby serving as a “don’t eat me” anti-phagocytic signal (132, 133). This combination also inhibited macrophage-mediated programmed cell removal and elimination of red blood cells that were mediated by complement receptors; this served to prevent the immune system from clearing normal cells (131, 134). Moreover, the binding of CD47 to SIRP-α inhibited integrin activation, thereby limiting macrophage proliferation in the vicinity of phagocytic targets and ultimately inhibiting phagocytosis (135). In addition, the regulation of macrophage phagocytosis is governed by the ratio of the activating ligand FcR to the inhibitory ligand SIRP-α. Fc receptor (FcγR) activation is inhibited upon binding of CD47 to SIRP-α, facilitated by dephosphorylated immunoreceptor tyrosine-based activation motifs (ITAMs) located adjacent after recruiting SHP-1 (136). Additionally, CD47 was expressed within macrophages, where it bound to SIRP-α expressed on the same macrophage. Inhibition of this cis interaction can induce hyper phagocytosis of macrophages (137).

CD47/SIRP-α axis is a well-established innate immune checkpoint. CD47 is often overexpressed on tumor cells and binds to SIRP-α expressed on macrophages to deliver the “don’t eat me” signal to macrophages, which ultimately results in the evasion of immune surveillance; this phenomenon has been correlated with poor prognosis (138–140). Direct blockade of CD47 on tumor cells via the binding of the Fc segment of CD47 IgG1 antibody to FcγR on macrophages neutralized the inhibitory CD47 signaling and activated antibody-dependent cell-mediated cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC), which has a significant inhibitory effect on tumor cells, such as leukemic cells in non-human primate (141). Several clinical trials and preclinical studies investigating drugs that promote phagocytosis by blocking the CD47 axis are currently ongoing or have been concluded. Relevant clinical trials are summarized in Table 3 . Nevertheless, using CD47 antibodies to treat tumors presents a challenge. Given the widespread expression of CD47, certain anti-CD47 antibodies exhibited similar targeting of normal cells, especially erythrocytes, leading to both normal cell-targeting toxicity and apparent antigenic silencing. Hu5F9-G4 and TTI-621 have been reported to induce adverse effects, such as anemia and thrombocytopenia (142, 143). Although it is possible to mitigate the adverse effects of Hu5F9-G4 by modifying the dosing regimen, it remains unclear whether this approach has an impact on efficacy (143). Hematological toxicity has been suggested to be Fc segment-dependent and this toxicity is due to SIRP-α-Fc fusion protein, whereas SIRP-α high-affinity monomers hardly cause any adverse effects (144–146). This suggests that optimizing the Fc structure of the anti-CD47 antibody may ameliorate these adverse effects. Regarding antigen silencing, the “don’t eat me” signal can be blocked by targeting the more restricted expression of SIRP-α. Therefore, screening antibodies or SIRP-α fusion proteins that weakly bind or do not bind to red blood cells using SIRP-α inhibitors or antibodies and the development of bispecific antibodies are crucial research avenues for future advancements in the field ( Figure 4 ).

The ATP-adenosine pathway is a crucial regulator of innate and adaptive immunity in the TME. CD39 and CD73 can catalyze the hydrolysis of ATP to adenosine. CD39, also known as ectonucleoside triphosphate diphosphohydrolase 1, converts extracellular ATP and ADP into AMP. Subsequently, CD73, an ecto-5’-nucleotidase, hydrolyzes AMP to adenosine. As reported previously, high extracellular ATP levels activate infiltrating inflammatory cells and promote the formation of an anti-tumor immune environment (157), whereas production of adenosine through hydrolysis is associated with the formation of an immunosuppressive microenvironment, such as the promotion of immunosuppressive cell infiltration (158).

CD39 was expressed on endothelial cells, tumor cells, and immune cells such as macrophages and DCs (159–161). Moreover, CD39 was expressed in >90% of B cells and monocytes in TME (162). TAMs with high expression of CD39 that exhibit increased enzymatic activity have been observed in ovarian cancer (163). Overexpression of CD39 has been associated with shorter OS and poor prognosis using samples from HCC and chronic lymphocytic leukemia patients (164, 165). CD73 is often expressed in the bone marrow, tumors, endothelial cells, etc. In pancreatic cancer, ovarian cancer, and glioblastoma, patients with high CD73 expression exhibited a prognosis worse than that exhibited by patients with low CD73 expression (166–168). Furthermore, in pancreatic cancer, CD73 expression was significantly positively correlated with PD-L1 expression (168). In glioblastoma, a high density of CD68⁺ macrophages expressing CD73 has been observed through single cell RNA-sequencing (167). Adenosine regulated intracellular cyclic adenosine monophosphate (cAMP) levels through the A1, A2a, A2b, and A3 receptor-mediated signaling pathways, thereby affecting anti-tumor immune response (162). Adenosine binding to A2a receptors inhibited the proliferation and cytotoxic activity of effector T cells (169), impaired ADCC of macrophages, which served as a “don’t eat me” signal ( Figure 5 ), and may promoted the migration of macrophages to tumors (170). Accordingly, high expression of CD39 and CD73 was associated with the formation of a tumor immunosuppressive microenvironment, which ultimately results in an impaired anti-tumor immune response.

CD39/CD73 expression on macrophages has profound impacts on anti-tumor immunity. Kynurenine produced by glioblastoma cells can activate aryl hydrocarbon receptors and upregulate the expression of CCR2 and CD39 on macrophages, thereby increasing adenosine levels and suppressing tumor-specific T cell response (171). IL-27, which is secreted by tumor-infiltrating neutrophils, has been reported to drive CD39 expression in macrophages to maintain a suppressive phenotype and increase IL-10 secretion (163). cAMP drove cAMP-response the transcription of Cd39, inducing its transcription by stimulating protein kinase A; this phenomenon may partially explain the mechanism underlying the upregulation of CD39 expression in TAMs (172). CircTMEM181 in HCC-derived exosomes sponged miR-488-3p in macrophages and upregulated CD39 expression on macrophages, activated the adenosine pathway, and induced M2-like polarization of the macrophages, thereby impairing anti-tumor immunity (173). Ludwig et al. demonstrated that exosomes derived from HNSCC cells were enriched in CD39, CD73 and adenosine, induced macrophages to transition to the M2 phenotype and promoted angiogenesis (174). HNSCC-derived small extracellular vesicles (sEVs) enriched in CD73 promoted CD73 expression on macrophages, triggered their polarization towards the M2 phenotype by activating the NF-κB pathway, simultaneously upregulating the expression of PD-1 and PD-L1 on macrophages. Depletion of CD73 in sEVs reversed resistance to PD-1 antibody therapy (175). These findings provide valuable insights into the regulation of CD39/CD73 on cell surfaces. Various drugs and antibodies have been developed to target CD39 and CD73. TTX-030 is a potent metamorphic inhibitor of CD39 activity and exhibited a maximum CD39 ATPase inhibition rate of 85% (176) and IPH5201 inhibited adenosine production by approximately 70%. In preclinical mouse models, the combination of IPH5201 and oxaliplatin, a chemotherapeutic agent that induced ATP production, significantly inhibited melanoma growth (177). Both drugs are currently undergoing clinical trials (NCT03884556, NCT04306900, and NCT04261075). Polyoxotungstate, an inhibitor of CD39, has been shown to increase the levels of M1 anti-tumor macrophages as well as CD8⁺ T cells and significantly inhibit the growth of CT26 colon cancer (178). AB680, a selective CD73 inhibitor, induced the proliferation and cytotoxic activity of depleted CD8⁺ T cells and significantly inhibited the growth of colon cancer established by CT26 cells; moreover, it exhibited favorable pharmacokinetic properties. Three phase I clinical trials (NCT04104672, NCT04575311, and NCT03677973) have been designed, two of which were recently concluded; however, their results are yet to be published (179). In addition, several studies have reported that CD39 or CD73 inhibitors and PD-1/PD-L1 mAbs exhibit synergistic anti-tumor effects. Zhang et al. modified an anti-CD39 IgG2 antibody by glycosylation that could deplete CD39-highly expressed immunosuppressive cells by enhancing antibody-dependent cytotoxicity; moreover, the modified anti-CD39 IgG2 and anti-PD-L1 immunotherapy exhibited favorable synergistic effects in mice models of melanoma and colorectal cancer (180). Kashyap et al. designed antisense oligonucleotides to efficiently knock down CD39 in regulatory T cells and TAMs in a mice model of breast cancer, thereby improving the response to anti-PD-1 antibody by upregulating PD-1 expression on CD8⁺ T cells (181). Li et al. reported that B66, an anti-CD39 antibody, reversed the resistance of colon cancer to PD-1 mAbs and the release of IL-18/IL-1β to inhibit colon cancer growth in mice by activating neutrophilic alkaline phosphatase 3 inflammatory vesicles, thereby elucidating another anti-tumor mechanism of anti-CD39 antibodies (182).

Targeting CD39 with therapeutic antibodies to enhance extracellular ATP signaling is a promising approach for cancer treatment and is expected to produce synergistic effects, ultimately resulting in improved tumor suppression when used in combination with chemotherapeutic agents that induce ATP release or PD-1/PD-L1 checkpoint inhibitors. In the field of antitumor immunity, CD73 or CD39 inhibitors have emerged as novel therapeutic targets that exhibit superior anti-tumor activity, improve immunosuppressive microenvironment, and demonstrate synergistic effects in combination with chemotherapy and immunotherapy in preclinical studies. These inhibitors are promising research prospects. Subsequent clinical trials are progressing to the preclinical phase, with concurrent safety assessments.

VISTA, a member of the B7 family, is highly homologous to PD-1. Its ligands are primarily V-set and Ig domain-containing 3 (183) and P-selectin glycoprotein ligand-1 (184). VISTA serves as a receptor and possesses an SHP-2 binding motif, as well as multiple casein kinase 2 and phosphokinase C phosphorylation sites in the cytoplasmic structural domain. It can signal VISTA-expressing cells in a PD-1-like manner when bound to its ligand (185). Hou et al. reported that VISTA expression is higher in immune cells than in pancreatic cancer cells and endothelial cells; moreover, among immune cells in melanoma specimens from patients, the number of CD68⁺ macrophages were higher than that of CD3⁺ T and CD19⁺ B cells (186). Patients with PD-L1-positive bladder cancer exhibited higher VISTA expression than patients with PD-L1-negative bladder cancer (187, 188). Furthermore, He et al. reported a significant correlation between PD-L1 and VISTA expression, which could be synergistically employed to predict prognosis in lymphoma (189). These findings indicate that the VISTA signaling pathway has the potential as a promising target for addressing the resistance encountered in currently available immune checkpoint therapies.

Siglec-10, an inhibitory receptor of the immunoglobulin superfamily containing two inhibitory motifs based on the cytoplasmic immune receptor tyrosine, is widely expressed on immune cells, such as macrophages and T cells, albeit primarily on macrophages (190, 191). CD24, an adhesion molecule, is a glycoprotein with multiple N- and O-linkage sites where sialic acid can be attached via α-2-3- or α-2-6-links. Siglec-10 recognizes and binds to sialic acid ligands with α-2-3 or α-2-6-links, thereby triggering an inhibitory signaling cascade mediated by SHP-1 and/or SHP-2 phosphatases. This is considered a “don’t eat me” signal on tumor cells to deter macrophage phagocytosis (192, 193). The enrichment of singles-10-positive cells within tumors has been associated with poor prognosis. TAMs with high siglec-10 expression exhibit a mixed M1/M2 phenotype and immunosuppressive activity and increased levels of immunosuppressive markers such as Arg-1, IL-10, TGF-β, TIM-3, and PD-L1. Paired with the PD-1 antibody, blocking siglec-10’s binding to CD24 improved the anti-HCC activity of CD8⁺ T cells and may synergistically promote tumor apoptosis which were from fresh human specimens with HCC patients (194). Targeting siglec-10 may be a promising approach to restoring anti-tumor immunity; however, further studies are required.

TAMs, which are immune cells that play participate in all stages of tumors, have attracted substantial interest because of their ability to influence tumor therapy. This review outlined the findings regarding the expression and regulation of various immune checkpoints on macrophages, their impact on prognosis, and their complex interactions with immunotherapy. Moreover, we discussed how these recent insights can be translated into anti-tumor treatment strategies and potential approaches to address the poor efficacy of immunotherapy. Currently, few of the clinical trials targeting the aforementioned immune checkpoints on macrophages have shown favorable therapeutic outcomes. The majority of these trials are still in the recruitment and ongoing phases, whereas few of them have been terminated prematurely owing to inadequate efficacy, needing more investigation.

The following research directions may be useful in the future to validate clinical trial outcomes: (1) improving the transition of preclinical research to clinical research, such as through the application of humanized animal models; (2) TAM subpopulations being isolated, characterized, and correlated with clinical outcomes precisely through the application of single-cell sequencing, multicolor immunofluorescence, spatial transcriptome, and machine learning, etc. and targeting key TAM subpopulations to enhance clinical therapeutic effects, as a promising strategy; (3) reprogramming TAMs to anti-tumor macrophages using TLR and CD40 agonists or PI3K-γ inhibitors to enhance immune activity as another feasible approach; (4) incorporating M1-like phenotypic genes using CAR vectors to construct engineered macrophages for tumor treatment; the potential of CAR-M to surmount host rejection and facilitate allogeneic therapy and overcoming the challenges of tumor immune escape and enhancing the yield of CAR-M; (5) the utilization of nano-delivery systems to precisely and efficiently target TAMs; however, the safety, tolerability, and drug delivery efficiency of nanoparticles requiring further evaluation.

Further and more comprehensive studies are required to clarify the indications for implementing these strategies and their efficacy and potential adverse effects on overall anti-tumor immunity. We anticipate that enhancing the efficacy of checkpoint inhibitors through the implantation of suitable strategies that improve the blockade of immune checkpoints expressed in TAMs will be essential for improving the therapeutic efficacy of existing immunotherapies and developing novel immunotherapeutic strategies.

All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.