Crosstalk between tumor-associated macrophages and the B7/CD28 family in immune checkpoint inhibitor-induced immunotherapy.

Introduction
Circulating monocytes derived from bone marrow are the precursors of macrophages. During tumorigenesis, monocytes are recruited to the tumor site by chemokines, such as chemokine ligand 2 (CCL2), colony-stimulating factor-1 (CSF-1), and subsequently differentiate into macrophages [1]. Macrophages can be classified as two major phenotypes: M1 (classically activated) phenotype, which was induced by lipopolysaccharides (LPS) and interferon-γ (IFN-γ), secretes pro-inflammatory cytokines such as interleukin-1β (IL-1β), IL-6, and IL-12 and plays antitumor role; M2 (alternatively activated) phenotype, which was induced by IL-4 and IL-13, secretes anti-inflammatory cytokines, such as transforming growth factor-beta (TGF-β), IL-10, and plays protumor role [2, 3]. However, macrophages differentiation is a dynamically changing process, and macrophages can exhibit intermediate or mixed states [4]. In the early stage of tumors, phenotype of macrophages is mixed and have both M1 and M2 types. M1 TAMs are dominant, initiate an inflammatory response and exert an antitumor effect [5]. However, once tumors are introduced, with tumor progression, cytokines and metabolites secreted by tumor cells reprogram TAMs from M1 subtype to M2 subtype, thus promoting angiogenesis, matrix remodeling, and suppressing antitumor immunity by increasing the expression of immune checkpoint molecules in tumor cells and other immunosuppressive cells [6–8].
Over the last decade, immunotherapies, especially immune checkpoint inhibitors (ICIs), have been the main breakthrough and have become promising therapeutic strategy for patients with a variety of cancers [9]. ICI-induced immune responses need to be strictly controlled by B7 and CD28 family members, which include immune checkpoint receptors and their corresponding ligands [10]. The following B7/CD28 family members have been extensively studied in the past decades: PD-1/PD-L1, CTLA-4/CD80/CD86, ICOS, B7-H3, B7-H4, VISTA, and HHLA2 [11]. B7/CD28 costimulatory/coinhibitory family molecules are expressed on macrophages and tumor cells and are prominent regulators of immune system. ICIs targeting B7/CD28 family members have gained much attention and were explored as treatments for several cancer types [12].
While current ICI treatment alone is effective in very limited cancer patients, a significant portion of patients do not respond to these ICIs (primary resistance), in part because of intratumoral heterogeneity and immunosuppressive tumor microenvironment (TME) [13]. In addition, a major subset of patients initially respond to ICI therapy, but inevitably develop resistance to therapy over time and eventually experience disease progression or recurrence (acquired resistance) [14, 15]. Drug resistance has been a massive obstacle in ICI treatment and hinders the clinical application of these drugs. Uncovering the resistance mechanisms is critical to enhance the efficacy of ICIs. Immunosuppressive TME is largely attributed to immunosuppressive cells including TAMs, myeloid-derived suppressor cells (MDSCs), and regulatory T cells (Tregs) [16]. Recent studies indicated that TAMs generally exhibit M2 phenotype. M2 TAMs are representative immunosuppressive cells in the TME. Polarizing the macrophage compartment from M2 phenotype to M1 phenotype can reprogram the TME to an immunostimulatory state. Furthermore, M2 TAMs are considered to be one of the main drivers of ICI resistance [17–19]. As the most studied immune checkpoints, B7/CD28 family members are supposed to promote TAMs polarization into M2 phenotype. M2 TAMs recruit regulatory Tregs through a cytokine network, thus inducing immunosuppressive TME and inhibiting antitumor immunity. In the process of ICI-induced immunotherapy, M2 TAMs can inhibit the activity and cytotoxic function of T cells by promoting the expression of immune checkpoint molecules in the TME, thus decreasing therapeutic efficacy [20]. In addition, TAMs, especially the M2 macrophages, often express immune checkpoint ligands to deliver inhibitory signals, suppress T cell activation, and block ICI-induced immune efficacy [21, 22].
However, TAMs are highly plastic cells that can be reprogrammed into either M1 or M2 with various microenvironmental stimuli [23, 24]. Single-cell RNA sequencing identified different TAM subpopulations with distinct functions and revealed TAM heterogeneity [25]. Thus, identifying the key modulators of monocyte recruitment and TAM polarization is considered to be a potent strategy to reverse TAM-mediated ICI resistance [26, 27]. There are several macrophage-targeting strategies and agents to reverse the tumor-promoting effect of TAMs: (1) inhibiting monocyte recruitment to the tumor site, such as CCL2/CCR2 antagonists and CXCR4 antagonists, (2) depleting M2 TAMs from tumor tissues, such as clodronate-loaded liposomes and zoledronate, (3) reprogramming TAMs from protumoral M2 to antitumoral M1 macrophages, such as CSF-1/CSF-1R inhibitors and PI3K-γ inhibitors, (4) inhibiting the secretion of immunosuppressive chemokines secreted by M2 TAMs, such as IL-10, TGF-β and CXCL2, (5) intervening with exosome sections from M2 TAMs, such as microRNAs, (6) targeting TAM-mediated immunosuppression, such as PD-1/PD-L1 blockade on TAMs and CD47/SIRPα blockade, (7) combination strategies to enhance efficacy, such as TAM-targeted therapy + chemotherapy/radiotherapy/ICI [28–32].
As shown in Fig. 1, in this review, we summarize the relationship between the B7/CD28 family members and TAM reprogramming, highlight the role of TAM-mediated immune escape in ICI immunotherapy, and explore the application prospects of TAMs in reversing ICI resistance.

Relationship between TAM polarization and PD-1, PD-L1
Programmed cell death protein-1 (PD-1, CD279) and programmed death-ligand 1 (PD-L1, B7-H1, CD274), the most classical coinhibitory signals, have been key targets in multiple solid tumors, and antibodies against PD-1/PD-L1 have shown remarkable therapeutic efficacy [33]. Drug resistance is a crucial factor that determines the efficacy of anti-PD-1/PD-L1 ICIs. In the TME, PD-L1 is highly expressed on tumors and binds to PD-1 on activated T cells to inhibit T cell activity, ultimately causing tumor immune escape. In addition to its expression on tumor cells, PD-L1 has also been found to be expressed in multiple types of antigen-presenting cells (APCs), especially in TAMs [34, 35]. TAMs and the PD-1/PD-L1 axis are key players in TME. Their interaction shapes the immunosuppressive TME and influences ICI efficacy. M2 TAMs express high PD-L1 through various signaling pathways, such as NF-kB, β-catenin, and PI3K/AKT/mTOR [13]. In addition, M2 TAMs secrete immunosuppressive cytokines such as IL-10 and TGF-β, which can upregulate PD-1 on exhausted CD8 + T cells. Furthermore, PD-1 is expressed not only on activated T cells, but also on TAMs in both human and mouse models, and PD-1 + TAMs are more likely to exhibit the M2 phenotype with increased CD206 expression [19]. Thus, TAMs may play a crucial role in regulating the PD-1/PD-L1 network.
In diagnosed cancers, the occurrence of drug resistance to PD-1/PD-L1 immunotherapy is involved with multiple complex mechanisms, such as genetic mutations, impaired antigen presentation, dysregulation of immune metabolism, and immunosuppressive TME [36, 37]. Arlauckas et al. indicated that PD-1/PD-L1 resistance is often closely related to the accumulation of intratumoral inhibitory macrophages [38]. On the one hand, PD-L1 expression of TAMs promoted the polarization of TAMs to M2 macrophages [39]. On the other hand, PD-L1 is expressed mainly on M2 TAMs rather than M1 TAMs, and PD-L1 + M2 TAMs accumulate over time with the progression of tumor [40]. Therefore, TAMs may be important targets for reversing resistance to anti-PD-1/PD-L1 therapy. Targeting TAM polarization can provide potent strategies to enhance the efficacy of PD-1/PD-L1 inhibitor therapy. Many investigations demonstrated that PD-L1 blockade promotes the M1 polarization of macrophages [41, 42].
A systematic review with 2277 patients suggested that there is a positive correlation between the expression of PD-L1 and TAM infiltration [43]. Patients with higher levels of PD-1 + TAMs were more likely to experience disease progression and early recurrence, which suggested that TAMs could act as prognostic biomarkers and therapy targets of PD-1/PD-L1 inhibitors.
USP7, one of the deubiquitinating enzymes, was first investigated for its role on TAM reprogramming by Dai et al. [44]. USP7 is highly expressed in M2 but not M1 macrophages. USP7 deletion suppressed macrophage M2 polarization, promoting CD8 + T cells proliferation and activation in vitro. Using a subcutaneous Lewis tumor-bearing mouse model, they found that USP7 inhibitors could effectively delayed tumor growth in vivo. In addition, USP7 inhibitors promoted the infiltration of M1 TAMs and CD8 + T cells by reprogramming TAMs from M2 to M1 phenotype. USP7 inhibitors increased the expression of PD-L1 in tumors, and the combination of USP7 and PD-1 inhibitors has synergistic antitumor effects.
Anti-PD-1/PD-L1 ICIs have shown very limited efficacy in the treatment of EGFR-mutated non-small cell lung cancer (NSCLC) because of the immunosuppressive TME created by activated EGFR signaling [45, 46]. Chen et al. demonstrated that immunoglobulin-like transcript (ILT) 4, an inhibitory receptor of the immunoglobulin superfamily, was upregulated by activation of EGFR-AKT/ERK1/2 signaling in NSCLC [47]. EGFR TKIs or specific inhibitors of AKT/ERK decreased the expression of ILT4 in EGFR-mutated NSCLC cells. By inducing the secretion of chemokines, such as CCL2 and CCL5, overexpressing ILT4 in EGFR-activated tumor cells induced the recruitment and M2-like polarization of TAMs, which impeded TAM-cell-mediated immunity. Anti-ILT4 can reverse TAM-induced immunosuppression and have synergistic antitumor effect with anti-PD-L1 agents. In a subcutaneous tumor model in C56BL/6 mice, ILT4 and PD-L1 combined blockade significantly reduced the proportions of CD206 + and CD163 + TAMs but increased those of CD86 + and CD80 + TAMs in tumor tissues, which indicated that ILT4 increased the efficacy of the PD-L1 inhibitor, at least partially by reprogramming TAMs from the M2 phenotype to the M1 phenotype. ILT4 blockade might be an effective treatment strategy to overcome ICI resistance in EGFR mutant NSCLC.
In colorectal cancer (CRC), especially in low levels of microsatellite instability (MSI-L) and microsatellite stable (MSS) CRC, the efficacy of present ICIs is also limited [48, 49]. Research on combination therapy is essential to potentiate the response of ICIs in those patients with cold CRC. Regorafenib, a VEGFR2 inhibitor, is one of the standard salvage-line treatments achieved improved PFS and OS for patients with CRC. A phase Ib trial of regorafenib plus the anti-PD-1 nivolumab showed encouraging antitumor efficacy in patients with MSS metastatic CRC [50]. One of the possible reasons is that VEGFR2 inhibitors can inhibit the function of immunosuppressive (M2-like) TAMs [51].
In triple-negative breast cancer (TNBC), especially in PD-L1 negative subgroups, the existence of immunosuppressive TAMs is one of the most important resistance mechanisms of ICI therapy. Eganelisib, a small molecule PI3K-γ inhibitor, can act as TAM-reprogramming agent which reshapes the TME by reducing TAMs recruitment and polarizing TAMs from M2 phenotype to M1 phenotype, thus to enhance the activity of ICIs. In a single-arm, phase 2 MARIO-3 study, the combination of eganelisib with the anti-PD-L1 atezolizumab and nab-paclitaxel showed encouraging survival benefit [52].
Programmed death-ligand 2 (PD-L2, B7-H2, CD273), another ligand of PD-1, is also expressed on TAMs and plays a role in immunosuppression. Compared with PD-L1, PD-L2 has higher affinity for PD-1 but is more restricted in expression. Combined PD-L1/PD-L2 blockade may improve immunotherapy efficacy in resistant cancers. There are no approved PD-L2-specific drugs yet, but the application of anti-PD-L2 monoclonal antibodies is under investigation. I think it would be valuable to explore how TAM modulation affects PD-L2-mediated immunosuppression.

Relationship between TAM polarization and other B7/CD28 family members
PD-1 and PD-L1 are the most commonly studied immune checkpoint molecules, and anti-PD-1/PD-L1 monoclonal antibodies have shown promising success in many types of solid tumors. However, anti-PD-1 ICIs work in only a subset of patients, and the response to PD-1/PD-L1 blockade in most malignancies is limited. In some tumors, other immune checkpoint molecules of the B7/CD28 family such as CTLA-4, B7-H3, and HHLA2, are more commonly expressed than PD-L1. Therefore, further investigations of the relationship between other B7/CD28 family members and TAMs are highly important for cancer treatment.

CTLA-4, CD80, CD86
Cytotoxic T lymphocyte-associated protein-4 (CTLA-4) is an immune checkpoint coinhibitory receptor expressed on activated T cells. B7-1 (CD80) and B7-2 (CD86) are its responding ligands [53]. Recent studies demonstrated that TAMs can also express CTLA-4 and CD80/CD86, suppressing T cell activity and contributing to immune evasion [54]. Ipilimumab antagonizes the binding of CTLA-4 to CD80/CD86 and was the first FDA-approved ICI in 2011 for patients with advanced melanoma, marking a milestone for advanced melanoma patients [55].
Recent studies have suggested that CTLA-4 may also be expressed on TAMs, contributing to shape immunosuppressive TME by promoting the M2 polarization of macrophages. Thus, targeting CTLA-4 on TAMs could also help to enhance antitumor immunity by reprogramming TAMs toward M1 phenotype [56]. Another study has showed that targeting sialoglycan in TME can promote M1 polarization of TAMs and enhance the treatment effect of CTLA-4 inhibitors [57]. Interestingly, propranolol, which was used for cardiovascular conditions, was found to improve the efficacy of anti-CTLA-4 by neutralizing the M2 polarizing effect [58].

ICOSL
Inducible T cell costimulatory ligand (ICOSL, B7-H2, CD275) is expressed on APCs, including macrophages. ICOSL binds ICOS on activated T cells. ICOSL + macrophages play a dual role in cancer immunity, either pro-inflammatory effects or immunosuppressive effects [59]. The combination of ICOS agonist and anti-CTLA-4 enhanced antitumor efficacy by remodeling TAMs to M1 phenotype [60].

B7-H3
B7-H3 is one of the most important immune checkpoints of the B7 family [61]. The human B7-H3 gene (CD276) is located on chromosome 15q24.1. B7-H3 is a type I transmembrane protein and is abnormally expressed on tumor cells and can be induced to be expressed on APCs, including dendritic cells (DCs) and macrophages [62]. To date, the B7-H3 receptor remains ambiguous. B7-H3 is overexpressed in numerous types of solid cancers and is often associated with an unfavorable prognosis [63, 64]. B7-H3 induces tumor angiogenesis, promotes hypoxia, dampens CD8 + T cell infiltration, and results in therapy resistance [65–68]. In the TME, B7-H3 controls immunity by inducing immunosuppression, including impairing T cell response, promoting M1 to M2 polarization of TAMs, and inhibiting the function of natural killer (NK) cells and DCs [69–71].
On the one hand, B7-H3-expressing tumor cells interact with receptors on TAMs to inhibit T cell activation. On the other hand, TAMs can express B7-H3 themselves to maintain an immunosuppressive TME [72]. Several studies have investigated the relationship between B7-H3 expression and TAM infiltration. Cheng et al. first established a correlation between B7-H3high TAMs and TNBC progression [73]. They reported that B7-H3 is enriched on TAMs in TNBC patients and tightly correlated with a poor clinical prognosis. B7-H3high TAM exhibit a mixture of M1/M2 phenotypes and exhibit immunosuppressive functions. Mao et al. found that the B7-H3 expression was positively associated with the CD68 + macrophage infiltration in CRC tissues and was negatively associated with survival rate of patients [74]. The B7-H3 signal could promote M2 macrophage polarization. Kang et al. reported that B7-H3 expression in tumor cells was correlated with TAM infiltration in Hepatocellular Carcinoma (HCC) tissues and predicts poor outcome. HepG2-secreted B7-H3 functioned in TAM regulation and could promote the M2 polarization of PMA-induced THP-1 cells partial via the STAT3 signaling pathway [75]. B7-H3 signal pathway may be involved in switching macrophages to the M2 phenotype by increasing the levels of IL-10 secretion. Patients with ovarian cancer respond poorly to PD-1/PD-L1 inhibitors. Miyamoto et al. reported a positive correlation between B7-H3 expression and the number of M2 macrophages in high-grade serous ovarian cancer (HGSOC) patients, especially in PD-L1-low non-immunoreactive phenotypes [76]. Mechanism research revealed that B7-H3 contributes to CCL2-CCR2-M2 macrophage axis-mediated immunosuppression, resulting in immune resistance. All above suggested that B7-H3 may interact with TAMs to form an immunosuppressive TME, may be a promising therapeutic target in cancer immunotherapy, and overcome tolerance.

B7-H4
B7-H4 (VTCN1, B7S1, B7x), a novel member of the B7 protein family, has been reported to be widely expressed in multiple types of cancer tissues [77, 78]. The human B7-H4 gene is located on chromosome 1p13.1. It is overexpressed on cancer cells and TAMs, contributing to enhance their immunosuppressive phenotype [79]. However, the receptor for B7-H4 remains to be identified.
B7-H4-expressing tumor cells secrete some cytokines to recruit monocytes to tumor site and then activate some signaling pathways to promote M2 polarization of macrophages. In a study of infiltrating ductal carcinoma, M2 macrophages present significantly higher expression levels of B7-H4 than M1 subtype [80]. Chi et al. reported that B7-H4 is a robust modulator of the M1/M2 macrophage polarization [81]. B7-H4 silencing in oral squamous cell carcinoma (OSCC) cell lines promoted M1 macrophage polarization and inhibited M2 polarization. In mice bearing OSCC tumor xenograft, B7-H4 silencing also repressed tumor growth and increased the M1/M2 ratio. Targeting B7-H4-TAMs axis may improve immunotherapy outcomes.

VISTA
V-domain Ig-containing suppressor of T cell activation (VISTA), also known as B7-H5, PD-1H, is an immune checkpoint molecule belonging to B7 family. VISTA is a type I transmembrane protein and mainly expressed on lymphocytes and myeloid cells [82]. Unlike other B7 family members, VISTA can serve as both a ligand and a receptor [83]. VISTA-targeting antibodies can regulate immunity response by promoting T cell infiltration, thereby reducing tumor growth in mouse cancer models. The VISTA signaling pathway has become a promising target for overcoming resistance to current ICIs [84].
In some cancers, VISTA signaling may drive macrophages toward immunosuppressive M2 phenotype. Jin et al. revealed the effect of VISTA on macrophage polarization in breast cancer and revealed that VISTA stimulates tumor growth by promoting M2 macrophage polarization through the STAT signaling pathway [85]. Lin et al. identified that ectopic expression of VISTA drives monocytes toward an M2 macrophage phenotype and regulates macrophage-mediated immune suppression [86]. VISTA expression is much greater in M2 macrophages than in M0 or M1 macrophages, and VISTA-expressing macrophages produced protumor cytokines, such as IL-10. These findings support the immunoregulatory role of VISTA in TAMs in human cancers and indicate that targeting VISTA on TAMs may be a potent strategy of cancer immunotherapy.

HHLA2
Human endogenous retrovirus-H long terminal repeat-associating protein 2 (HHLA2, also known as B7-H7, B7y) is a recently discovered member of the B7 family and has attracted great attention in immunotherapy [12, 87]. The B7-H7 gene is composed of an open reading frame (ORF) containing 414 amino acids and shares 23–33% amino acid identity with the other B7 family members [88]. The HHLA2 protein is overexpressed in various human cancers, especially in PD-L1-negative cancers, but is absent in mice and rats. HHLA2 can act as both a costimulatory and coinhibitory signal depending on the different receptor it binds [89]. HHLA2 promotes/inhibits the proliferation and cytokine secretion of CD4 + and CD8 + T cells by binding to its costimulatory receptor TMIGD2 and its coinhibitory receptor KIR3DL3 [90–93].
In the human immune system, HHLA2 is detected mainly on monocytes/macrophages. Han et al. demonstrated that higher expression of HHLA2 in CD68 + TAMs was significantly correlated with poorer OS of pancreatic cancer patients and could serve as an important biomarker for the prognostic prediction of the patients [94]. Wang et al. found that high HHLA2 expression is related to increased macrophage infiltration and induced M2 polarization of macrophages. Tumor-related M2 macrophages can induce tumor-promoting cytokine secretion, in turn increasing the expression of PD-L1 [95]. In addition, HHLA2 promotes immune tolerance by activating the JAK/STAT signaling pathway in HCC cells [96]. Moreover, our previous study found that in co-culture model of NSCLC cells and THP-1 macrophages, HHLA2 knockdown suppressed M2 polarization of TAMs by reducing the expression and secretion of IL-10 [97]. Overall, blocking HHLA2 or its receptors may reverse TAM-mediated immunosuppression and improve immunotherapy responses.
In many tumors, especially in PD-L1-negative tumors, these lesser-discussed checkpoints also play critical roles in TAM polarization and contribute to immunosuppression. More relative studies are going on.

TAM-reprogramming agents combined with ICIs in cancer treatment
From what mentioned above, we can see that there is a complicated interplay between TAMs and B7/CD28 family members in ICI resistance. Therefore, therapeutic strategies targeting TAMs may offer an opportunity to enhance response to ICI in cancer patients. On the one hand, TAM-reprogramming agents can reduce the suppressive macrophage burden, thus to maximize the antitumor activity mediated by ICIs. On the other hand, ICIs can reverse the polarization of M2 TAMs and increase the proportion of pro-inflammatory M1 TAMs, which in turn increase the effect of ICIs.
Recently, numerous preclinical and clinical researches are focused on exploring treatment strategies targeting TAMs combined with ICIs to overcome TAM-induced ICI resistance and improve the therapeutic efficacy, and some trials have already shown promising results, as shown in Table 1. CSF-1/CSF-1R inhibitors can remodel TME by altering TAM phenotype and are one of the most used TAM-target agents. PLX3397 is an FDA-approved CSF-1/CSF-1R inhibitors for the treatment of tenosynovial giant cell tumor in August 2019 [98]. The combination of PLX3397 and αPD-1/αCTLA-4 regressed tumor burden more than 85% in pancreatic ductal adenocarcinoma mouse model by reducing immunosuppressive activity of TAMs [99]. In MC38 and B16 models, M2-like macrophages impeded the anti-PD-1 response, and tumor growth were significantly inhibited in PLX3397 + anti‐PD‐1 group [100]. C19, a novel CSF-1R inhibitor, can efficiently repolarize M2-Like TAMs to M1 phenotype and improves antitumor efficacy of PD-1 antibody on subcutaneous MC38 models [101]. Apart from CSF-1R inhibitor, there are also other several TAM-target agents applied to preclinical research to enhance the efficacy of ICIs, such as CCR2 antagonist, PI3K inhibitor, USP7 inhibitor, and anti-TGF-β [44, 102, 103]. The specific mechanisms of those TAM-target agents in reversing ICI resistance are shown in Fig. 2.

While preclinical research has made great progress, our ultimate target is to translate these advances into clinic and improve the outcomes of cancer patients, the related clinical research is ongoing, as shown in Table 2. We are looking forward to the results.

Future perspectives
The application of ICIs has made considerable progresses in cancer therapeutics, and compared to standard approaches, ICIs improve patient prognosis and induce fewer adverse events. ICIs are currently recommended as the first-line management in many types of malignant cancers, such as NSCLC, urothelial carcinoma, and melanoma [111, 112]. While low response rate of monotherapy with ICIs restricted its application in clinic. In addition, natural or acquired resistance to ICIs greatly limits their therapeutic potential and represents one of the main challenges in medical oncology.
The effect of ICIs is closely related to the TME. Macrophages account for the largest proportion of tumor infiltrating immune cells [113, 114]. TAMs typically exhibit an M2-like phenotype, which can secret immunosuppressive chemokines and cytokines and is associated with tumorigenic properties, whereas M1 macrophages promote antitumor immunity and exert antitumor effects. Therefore, reprogramming TAMs into M1-like macrophages would help to reprogram the immunosuppressive TME toward an immune reactive state and would be an effective approach to eliminate immune suppression and promote tumor regression. An increasing amount of evidence suggests that TAM-reprogramming therapy, as a new but rapidly developed strategy of cancer therapy, could effectively reverse resistance and improve the efficacy of immunotherapies, especially ICIs [115].
Currently, the main research focus is on designing strategies to identify new TAM-reprogramming strategies and agents for precision cancer medicine. There are some limitations in current research. Firstly, M1/M2 macrophage polarization system has been used to describe macrophage diversity in the past decades; however, the state of macrophages is complicated and not so obsolute. Through single-cell RNA sequencing technologies, TAMs can be divided into five even seven subsets with different phenotypes and functions [116–119]. Secondly, although TAMs are supposed to regulate ICI-induced immunoreaction and targeting TAMs may be a potent strategy to recover the efficacy of PD-1/PD-L1 inhibitors, the causal relationship between TAM polarization and ICI resistance in TME is complex and has not been thoroughly elucidated to date. Thirdly, some TAM-targeting agents work in certain cancers but not others, identifying subpopulations of patients likely to benefit from different therapies targeting macrophages is essential. Furthermore, selecting the optimal combination of TAM-targeting agent and ICI for certain types of tumors is difficult. Despite these difficulties, there is still great potential to explore new strategies of modulating macrophages to optimize the therapeutic efficacy of ICIs in future preclinical and clinical trials. Further investigations are needed, and future strategies may include: deciphering TAM heterogeneity, choosing target molecules, selecting patients who will benefit, optimizing combination strategies.
In conclusion, macrophages have been shown to influence the TME, and targeting macrophages might be a novel strategy to reprogram the immunosuppressive TME and overcome immunotherapy resistance. Hence, more approaches targeting macrophage reprogramming in combination with other immunotherapies need to be explored in the future.