The role of tumor-associated macrophages in osteosarcoma progression – therapeutic implications

Abstract

Background

Osteosarcoma (OS) is the most common primary malignant bone tumor. Compared with previous treatment modalities, such as amputation, more recent comprehensive treatment modalities based on neoadjuvant chemotherapy combined with limb salvage surgery have improved the survival rates of patients. Osteosarcoma treatment has, however, not further improved in recent years. Therefore, attention has shifted to the tumor microenvironment (TME) in which osteosarcoma cells are embedded. Therapeutic targets in the TME may be key to improving osteosarcoma treatment. Tumor-associated macrophages (TAMs) are the most common immune cells within the TME. TAMs in osteosarcoma may account for over 50% of the immune cells, and may play important roles in tumorigenesis, angiogenesis, immunosuppression, drug resistance and metastasis. Knowledge on the role of TAMs in the development, progression and treatment of osteosarcoma is gradually improving, although different or even opposing opinions still remain.

Conclusions

TAMs may participate in the malignant progression of osteosarcoma through self-polarization, the promotion of blood vessel and lymphatic vessel formation, immunosuppression, and drug resistance. Besides, various immune checkpoint proteins expressed on the surface of TAMs, such as PD-1 and CD47, provide the possibility of the application of immune checkpoint inhibitors. Several clinical trials have been carried out and/or are in progress. Mifamotide and the immune checkpoint inhibitor Camrelizumab were both found to be effective in prolonging progression-free survival. Thus, TAMs may serve as attractive therapeutic targets. Targeting TAMs as a complementary therapy is expected to improve the prognosis of osteosarcoma. Further efforts may be made to identify potential beneficiaries of TAM-targeted therapies.

Introduction

Osteosarcoma is the most common primary malignant bone tumor [1]. It occurs in the metaphysis of long bones, most notably around the knee joints [2–4]. The peak incidence is in children and adolescents aged 10–24 years [5, 6]. Although osteosarcoma is a relatively rare disease with an incidence of approximately three per million, its mortality rate was high before the advent of neoadjuvant chemotherapy, with a disease-free survival of less than 20% [7]. Compared to procedures used in the past, including amputation, more comprehensive therapy based on neoadjuvant chemotherapy combined with limb salvage treatment has effectively improved the survival time of patients with osteosarcoma and increased the survival rate from 20% to 60–70% [8, 9]. In recent years, however, osteosarcoma treatment has not further improved, regardless increases in dosages of neoadjuvant drugs or in the types of drugs used in chemotherapy regimens. The 5-year survival rate of patients with localized osteosarcoma is still no more than 70% [10], and the 3-year survival rate of patients with metastatic high-grade osteosarcoma is only about 20% [11]. The recurrence and metastasis rates remain around 30–40% and 15–20%, respectively. The lung is the most common site of metastasis, with a frequency up to 80% [12, 13]. To make matters worse, almost all patients who relapse develop metastases [14]. Therefore, making early diagnoses, preventing and limiting lung metastasis, improving treatment effects and reducing the acquisition of drug resistance are the major challenges faced by bone oncologists. It has increasingly become apparent that focusing only on osteosarcoma cells as targets is not likely to further improve the therapeutic results and, so, attention has gradually shifted to the TME, in which osteosarcoma cells are embedded.

Tumors do not exist in isolation, but instead are closely associated in time and space with their surrounding microenvironment, and even the entire body. So, the occurrence and development of osteosarcoma and its complex TME are inseparable [15]. The TME induces metabolites and other factors that promote the proliferation and metastasis of osteosarcoma cells, as well as the acquisition of drug resistance [16, 17]. It is a complex ecosystem composed of various components, including malignant cells, endothelial cells, tumor-infiltrating immune cells and stromal cells with different functions [18, 19]. Immune cells represent an important part of the TME [20]. Innate immune cells, such as macrophages and suppressor cells derived from bone marrow, and adaptive immune cells, such as T cells and B cells, have been reported to participate in the evolution of various tumors, including glioblastoma, tongue squamous cell carcinoma and esophageal cancer [21–24]. In the process of malignant transformation of cells, the TME provides the necessary support. Macrophages are abundantly recruited to tumor sites and distributed during different stages of tumor progression [25]. Therefore, it is generally believed that macrophages may play important roles in tumor progression [26–28]. Macrophages that infiltrate tumor tissues or accumulate in their microenvironments are defined as tumor-associated macrophages (TAMs) [29], including monocyte-derived macrophages (MDMs) and tissue-resident macrophages (TRMs) that are derived from yolk sac progenitor cells [30]. In recent years, researchers have begun to study the role of TAMs in osteosarcoma development, aiming at improving its treatment. However, there is currently no consensus on the exact role played by TAMs in the progression of osteosarcoma. In addition to the conventional polarization of TAMs into M1 and M2 phenotypes there are also TAMs with both phenotypes. In addition, TAMs may show dual effects of promoting and suppressing cancer, and may promote tumor angiogenesis without consequences for prognosis. Here, we will discuss the source of TAMs and their mechanism of action in the progression of osteosarcoma, the duality of TAMs and their potential as therapeutic targets.

Source and classification of macrophages

There are two main sources of macrophages: (1) MDMs derived from bone marrow, which are the precursors of macrophages in e.g. the intestine and heart [31, 32] and (2) TRMs of the yolk sac and embryonic liver [33], which are the precursors of e.g. brain microglia [34, 35], alveolar macrophages [36] and liver Kupffer cells [37] (Fig.1). According to the different pathways of activation, macrophages are divided into two main categories, i.e., classical activated macrophages (M1) and selective activated macrophages (M2). Among them, M1-type macrophages can be activated by lipopolysaccharide (LPS), interferon-γ (IFN-γ) or granulocyte-macrophage colony-stimulating factor (GM-CSF), and highly express interleukin (IL)-1, IL-6, IL-12, TNF-α and inducible nitric oxide synthase (NOS). They also exhibit pro-inflammatory, immunogenic and anti-tumorigenic properties [38, 39]. M2-type macrophages can be activated by factors such as IL-4 and IL-13 to produce transforming growth factor (TGF)-β, IL-10 and angiogenesis factors [40], which have anti-inflammatory and tumor-promoting effects [41] (Fig. 1). M2-type macrophages can be further divided into three subtypes: M2a, M2b and M2c [42, 43]. Among them, M2a macrophages are induced by Th2-derived cytokines IL-4 or IL-13, M2b macrophages by immune complexes, LPS, TLRs or IL-1 receptor antagonists (IL-1ra), and M2c macrophages by IL-10, TGF-β or glucocorticoids (GCs) [43]. Some researchers have pointed out that TAMs display phenotypes and functions similar to M2-type macrophages and, therefore, may constitute a group of M2 polarized macrophages [44, 45]. In a study on ovarian cancer macrophages, an additional M2d subtype has been proposed [46]. The authors found that leukemia inhibitory factor (LIF) and IL-6 could induce monocytes to consume monocyte colony stimulating factor (M-CSF), thereby allowing them to differentiate into TAMs that are distinct from M2a-c. In order to unify this concept, a group of macrophage experts at the International Congress of Immunology in Milan in August 2013 proposed a concept of macrophage activation spectrum based on stimulation conditions, thereby classifying macrophages as M(IL-4), M(IL-10), M(GC), M(IFN-γ), M(LPS) and M(LPS +IFN-γ) [47]. At present, however, the most widely used classification is restricted to M1 and M2.

Fig. 1.

Sources and classification of macrophages. Macrophages are mainly differentiated from hematopoietic stem cell-derived monocytes in the bone marrow. These macrophages enter the blood stream and become monocyte-derived macrophages (MDMs) after reaching the tissues. Other macrophages are derived from the yolk sac and embryonic liver. Some of the macrophages derived from the yolk sac appear before hematopoiesis in the liver and differentiate directly into tissue-resident macrophages (TRMs). Some of these macrophages are derived from yolk sac-derived myeloid-biased progenitors (YSMPs) that migrated to the embryonic liver, differentiated into monocytes, entered tissues and differentiated into TRMs. The macrophages that enter tumor tissues become TAMs and differentiate into M1-type or M2-type TAMs under the action of various cytokines

The mechanism of action of TAMs in the progression of osteosarcoma

The TME contains a variety of infiltrating immune cells, of which TAMs are the most abundant ones. The proportion of TAMs in osteosarcoma accounts for more than 50% of the immune cell content [48, 49]. In lung cancer, breast cancer and other cancers, the content of TAMs can even reach 50% of the tumor volume [50, 51]. TAMs present in tumor tissues can be supplemented by peripheral recruitment of TAMs derived from monocytes. In addition, some TAMs can be derived from the in situ proliferation of tissue macrophages [52, 53]. Tissue macrophages can also be induced by the TME to become TAMs [54]. In recent years, ample studies have reported on the role of TAMs in tumor growth and metastasis [55–58], including the vital role they play in the occurrence and development of osteosarcoma. The different roles of TAMs in osteosarcoma and their specific biomarkers are described below (see also Table 1).

Table 1.

Biomarkers and signaling pathway molecules associated with tumor-associated macrophages in osteosarcoma (OS)

Biomarkers	Functions in OS	M1	M2	Associated signaling pathway molecules	References
CD209	Tumorigenesis	–	+	/	[65]
COX-2	Lymphoangiogenesis, metastasis	/	/	COX-2/STAT3	[101]
IL-1β	Lymphoangiogenesis, drug resistant	+	–	VEGF	[101, 112]
CD163	Immunosuppression	–	+	/	[105]
CCL22, TGFB2	Immunosuppression	–	+	TGFB2	[110]
IL-10	Immunosuppression, metastasis	–	+	TGFB2, STAT3	[110, 114]
CCL18	Proliferation, migration	–	+	EP300/UCA1/Wnt/β-catenin	[113]
MCP-1, TGF-β1	Metastasis	–	+	STAT3	[114]

The role of different TAM polarization types in osteosarcoma

As pointed out above, the M1 and M2 subtypes represent the two extreme states of TAMs. The polarization process of TAMs depends on various signals within the TME, and the overall polarization of TAMs during tumor progression is conducive to tumor growth [59]. Under the action of various signals within the TME, M1-type and M2-type TAMs can transform into each other to a certain extent [60–62]. Therefore, it is believed that the ideal method to target TAMs is not to eliminate them, but to transform tumor-promoting M2-type TAMs into anti-tumor M1-type TAMs [63].

Using immunohistochemical staining, it was found that localized osteosarcoma tissues are highly infiltrated with M1-polarized TAMs, while metastatic osteosarcoma tissues contain M2-polarized TAMs. An imbalance in the ratio of M1-type and M2-type TAMs may affect the metastatic potential of osteosarcoma cells, whereas M1 polarization can inhibit osteosarcoma metastasis [64]. Another study showed that M2-type TAMs were enriched in primary osteosarcoma tissues. The authors successfully inhibited the colony and sphere-forming abilities of osteosarcoma cells and osteosarcoma formation in mice after inhibiting the M2 polarization of TAMs with the drug all-trans retinoic acid (ATRA) [65]. In addition, overexpression of long non-coding RNA RP11-361F15.2 promoted the M2 polarization of TAMs, thereby increasing the migration and invasion of osteosarcoma cells and promoting the growth of xenograft tumors in vivo [66]. In mouse osteosarcoma xenograft models it has been found that expression of the epidermal growth factor receptor (EGFR) can also mediate the tumor-promoting effect of TAMs [67], with M2 polarization at the inoculation site. Interestingly, it has also been found that photothermal therapy (PTT) can inhibit osteosarcoma progression, and that PTT can inhibit M2 polarization, thereby inhibiting the invasive ability of osteosarcoma cells in mice [68]. Thus, polarization of TAMs is common in osteosarcoma and is one of the main processes governing osteosarcoma progression.

Mifamurtide has been found to regulate osteosarcoma progression through M1/M2 polarization of TAMs [48]. Mifamurtide is a synthetic drug simulating bacterial cell wall components that stimulate immune responses and activate macrophages and monocytes [69]. In some European countries, Mifamurtide has been used clinically to improve the overall survival of patients with osteosarcoma [70]. It has been found that Mifamurtide can affect M1 and M2 polarization, resulting in an intermediate type of TAMs affecting the proliferation and migration of osteosarcoma cells [48]. As such, it is worth noting that it has been shown that TAMs in primary osteosarcomas exhibit both M1 and M2 characteristics. The authors suggested, however, that the prognosis of osteosarcoma may not be related to the M1 and M2 polarization of TAMs, but rather to the total number of TAMs [71]. This notion suggests that TAMs may be alternatively involved in the progression of osteosarcoma.

TAMs play a role in osteosarcoma metastasis through a pre-metastatic microenvironment

There is common consensus that tumor metastasis is not an exclusive advanced process, and that cellular changes take place before tumor cells reach metastatic sites. These sites, that are selected by the tumor cells, are prone to metastasis and are called pre-metastatic microenvironments (PMNs) [72, 73]. As early as 1889, Steven Paget put forward the “seed and soil” metastasis theory based on the analysis of 735 breast cancer cases, combined with other tumor related studies. The “seed” refers to certain tumor cells with metastatic potential, and the “soil” refers to any organ or tissue that provides a suitable environment for the growth of the “seed”. Paget believed that the location of tumor metastases is not random, and that the spread of tumor cells is organ-specific, with metastasis occurring only when the tumor cells are compatible with the local environment [74, 75]. Osteosarcoma has a clear tendency to metastasize to the lung [76, 77]. As an experimental model, osteosarcoma 143B cells stably transfected with green fluorescent protein (GFP) were transplanted into the tibia of nude mice transgenic for red fluorescent protein (RFP). Five weeks later, bone tumors expressing GFP with a matrix expressing RFP (including TAMs) were implanted into the tibia of non-transgenic nude mice. After another 6 weeks, metastases with green fluorescent osteosarcoma cells embedded in red fluorescent stromal cells were observed in the lungs of the non-fluorescent nude mice, and subsequent metastasis from the tibia to the lungs was observed using high-power confocal microscopy [78]. This study nicely demonstrated a close in vivo relationship between TAMs, osteosarcoma cell growth and lung metastasis, in which osteosarcoma cells that metastasized to the lung were surrounded by TAMs derived from the primary tumor. An additional study showed that M2-type TAMs can enhance lung metastasis of mouse osteosarcoma K7M2 WT cells. As pointed out above, the active derivative of vitamin A, ATRA, can inhibit metastasis by inhibiting M2 polarization of TAMs. In this study, ATRA not only reduced the number of osteosarcoma lung metastatic lymph nodes, but also reduced the infiltration of M2-type TAMs into metastatic lymph nodes [79]. Whereas M2-type TAMs may provide the necessary “soil” for the metastasis of osteosarcoma cells, ATRA may inhibit the M2 polarization of TAMs and cause “soil” changes, thereby inhibiting the metastasis of osteosarcoma cells. In addition, by using a lung metastasis model of osteosarcoma in mice, Han Yu et al. found that the density of M2-type TAMs in the lung metastases was higher than that in the primary tumors, and confirmed that M2-type TAMs increased the invasive ability of osteosarcoma cells [80]. Thus, TAMs may provide a suitable environment for lung metastasis of osteosarcoma.

The role of TAMs in the process of osteosarcoma stem cell transformation

Cancer stem cells (CSCs) are also called tumor-initiating cells or tumor sustaining cells, and constitute a stem cell-like subpopulation within the tumor cell mass [81, 82]. CSCs are highly resistant to chemotherapy and radiotherapy. Elimination of CSCs may reduce the drug resistance of tumors and, thus, prevent tumor recurrence [65]. It has been found that TAMs can directly interact with CSCs and maintain their stem cell-like characteristics, thereby triggering tumorigenesis and tumor progression [25, 83]. In addition, it has been found that CD209-positive M2-type TAMs can activate CSCs to promote osteosarcoma formation (Fig. 2), while ATRA can inhibit the colony forming and sphere-forming abilities of osteosarcoma cells in vitro and TAM-induced osteosarcoma formation in vivo in mice by reducing the activity of CSCs and inhibiting M2-type TAMs [65].

Fig. 2.

Implications of the main molecular protagonists related to TAMs in osteosarcoma. TAMs deplete tumor-infiltrating T cells and inhibit T cell proliferation and production of pro-inflammatory factors. The cytokines secreted by TAMs directly or indirectly promote the proliferation or metastasis of osteosarcoma. Osteosarcoma (OS) cells can escape phagocytosis by TAMs after binding of CD47 on its surface to SIRPα on TAMs. M2-type TAMs activate CSCs to promote the formation of osteosarcoma. Lymphangiogenesis in osteosarcoma is related to COX-2 expression and IL-1 production by TAMs. IL-34 secreted by TAMs promotes angiogenesis of osteosarcoma. IL-1β produced by TAMs mediates drug resistance of osteosarcoma

The role of TAMS in angiogenesis and lymphangiogenesis of osteosarcoma

Most solid tumors secrete angiogenic factors to induce blood vessel growth, thereby providing oxygen and nutrients to the tumors [84]. Tumor blood vessels are important for enabling tumor volume enlargement and distant metastasis [85, 86]. It has been found that the cytokine IL-34 can promote M2 phenotype polarization of osteosarcoma TAMs and the formation of osteosarcoma blood vessels (Fig.2), as well as induce the adhesion of monocytes to activated endothelial cells, leading to osteosarcoma metastasis [87]. Surprisingly, however, no correlation was noted between vascular density and the prognosis of osteosarcoma. Although M2-type TAMs can promote the vascular density of osteosarcoma, it does not seem to affect its prognosis [71].

Lymphatic vessels also play an important role in the metastasis of tumors to nearby or even distant lymph nodes through the lymphatic system [88–90]. Therefore, reducing the formation of lymphatic vessels may be a way to inhibit tumor metastasis. Stilbenes is a polyphenolic plant toxin [91], and its derivative resveratrol can inhibit the proliferation of breast cancer cells and other tumor cells [92, 93]. It has anti-oxidation, anti-inflammation, anti-apoptosis, anti-tumor, anti-angiogenesis, anti-lymphangiogenesis and other effects [94–98]. In osteosarcoma, it has been found that resveratrol treated conditioned medium of M2-type TAMs can inhibit the movement, migration and tube formation of human lymphatic endothelial cells induced by vascular endothelial growth factor (VEGF)-C (Fig. 2), and successfully promote the formation of lymphatic vessels in mice [99]. Similarly, two other derivatives of stilbenes, 2,3- and 4,4-dihydroxystilbene, have been found to inhibit the differentiation of M2-type TAMs by reducing the phosphorylation of STAT3 in tumor-bearing mice, thereby inhibiting the formation of lymphatic vessels, tumor growth and metastasis to lung and liver [100]. In addition, wogonin has been found to inhibit lymphangiogenesis by inhibiting COX-2 expression and IL-1 generation in TAMs and reducing VEGF-C-induced VEGFR-3 phosphorylation, thereby playing an anti-tumor and anti-metastatic role [101].

The role of TAMs in immunosuppression of osteosarcoma

At present, the efficacy of immunotherapy is limited due to various immunosuppressive signals within the TME [102], in which TAMs play an important role. It has been shown that TAMs can inhibit the recruitment of CD8+ T cells to the TME [103]. In addition, TAMs can inhibit the tumor-killing functions of cytotoxic T cells by consuming metabolites necessary for T cell proliferation [104]. In osteosarcoma, it has been found that stimulation of TIM-3(+) PD-1(-) and TIM-3(+) PD-1(+) T cell receptors reduced the proliferation of T cells and the secretion of pro-inflammatory cytokines. In addition, it was found that TAMs may play an important role in antigen presentation to T cells, and that the occurrence of CD163(+) M2-type TAMs was directly related to the frequency of TIM-3(+) PD-1(+) T cells and intensified immunosuppression of T cells [105] (Fig. 2).

The CD11b+ bone marrow subgroup represents an important part of the tumor immune microenvironment, including myeloid-derived suppressor cells (MDSCs) and TAMs [106], which play important roles in suppressing T cell-mediated anti-tumor immunity. They stimulate the differentiation of regulatory T cells (Tregs) and recruit Tregs to tumor tissues [107]. These cellular components are considered to be the main negative regulators of T cell-mediated anti-tumor immunity and to play an important role in suppressing T cell-mediated anti-tumor immunity [108]. It has, however, also been found that metformin can induce CD11b+ cells to exert a tumor suppressor effect independent of T cells, leading to growth inhibition of osteosarcoma. In addition, it was found that in wild-type (WT) mice and SCID mice with a depletion of T cells, growth inhibition of osteosarcoma can still be observed [106].

Exosomes are extracellular carriers with a diameter of 30–150 nm that have been found to be closely related to the progression of osteosarcoma [109]. It has also been found that incubating MHS mouse alveolar macrophages with exosomes of metastatic osteosarcoma cells results in induction of the expression of IL-10, TGF-β2 and CCL22, and reduction in TAM-mediated phagocytosis, exocytosis and tumor cell killing (Fig. 2). The exosomes of this metastatic osteosarcoma cell can regulate TAM cell signaling, thereby promoting their M2 polarization, leaving tumor growth and proliferation uncontrolled, and generating an immunosuppressive and tumorigenic microenvironment through TGF-β2 [110].

The role of TAMs in osteosarcoma drug resistance

Drug resistance is a challenge faced by many tumor chemotherapy regimens, and is an important factor determining the poor efficacy of neoadjuvant chemotherapy for osteosarcoma. In pancreatic ductal adenocarcinoma (PDA), TAMs can release deoxycytidine, which inhibits gemcitabine through molecular competition at the drug uptake and metabolism levels, leading to PDA resistance to gemcitabine [111]. Similarly, under stimulation of the neoadjuvant chemotherapeutic drug CDDP in osteosarcoma, TAMs can secrete IL-1β in large quantities (Fig. 2), thereby reducing the sensitivity of osteosarcoma cells to CDDP, leading to drug resistance [112]. Thus, TAM-mediated drug resistance of osteosarcoma cells may be an important cause of stagnation of neoadjuvant chemotherapy.

Other TAM-related mechanisms involved in osteosarcoma progression

In addition to the abovementioned common mechanisms of action, TAMs may also affect the progression of osteosarcoma in other ways. M2-type TAMs can for example secrete CCL18 and promote the proliferation and metastasis of osteosarcoma cells through the EP300/UCA1/Wnt/β-catenin pathway [113] (Fig. 2). Also, dihydroxycoumarins (i.e., esculetin and fraxetin) can inhibit the growth and metastasis of osteosarcoma cells to the lung or liver. It appears that a reduction in phosphorylation of STAT3 inhibits the production of IL-10, monocyte chemotactic protein (MCP)-1 and TGF-β1 during the differentiation of M2-type TAMs [114].

TAMs inhibit osteosarcoma progression: the duality of TAMs

In most studies, the density of TAMs has been found to be related to a poor prognosis of cancer patients [57, 115, 116], whereas in less than 10% of studies the density of TAMs in the TME was found to be associated with a good prognosis [117]. This duality has been noted in a variety of cancers, including prostate, lung and brain cancer [118]. Also in osteosarcoma a similar duality appears to be present. The first part of this review dealt with the cancer-promoting effects of M2-type TAMs in osteosarcoma, but some investigators found that TAMs can also inhibit the growth and metastasis of osteosarcoma, and be related to a good prognosis. It has been shown that both the M1 and M2 subtypes of TAMs can inhibit the growth of osteosarcoma cells under certain conditions. When stimulated for example by LPS + IFN-γ, M1 polarized TAMs can inhibit the growth of osteosarcoma cells, whereas IL-10 stimulated M2 polarized TAMs can inhibit the growth of osteosarcoma cells under the action of anti-EGFR cetuximab, which involves antibody-dependent tumor cell phagocytosis [69]. Several studies have indicated that TAMs in osteosarcoma may have both M1 and M2 characteristics and that the prognosis of osteosarcoma is not significantly related to the M1 and M2 subtypes of TAMs but, instead, is positively correlated with the total number of infiltrating TAMs. Patients with high TAM counts are less likely to develop metastases. Moreover, it has been found that the density of TAMs in primary localized osteosarcomas is about twice that of metastatic osteosarcomas, and that the presence of TAMs inhibits the metastasis of osteosarcoma cells [71].

Clinical trials aimed at targeting osteosarcoma macrophages and immune checkpoints

Based on the various important roles played by TAMs in osteosarcoma, several clinical trials have been carried out and/or are in progress (Table 2). Clinical trials using Mifamurtide are classic and numerous. Mifamurtide is a synthetic mimetic bacterial cell wall component that acts as an activator of macrophages. Under the induction of IFN-γ, Mifamurtide can activate macrophages to exert direct anti-tumor activities [69]. It has been shown that combinations of chemotherapy (cisplatin, doxorubicin, methotrexate, ifosfamide) with Mifamurtide can improve the overall survival and progression-free survival (PFS) of patients with osteosarcoma [119]. In patients with metastatic disease, however, Mifamurtide combined with chemotherapy did not significantly improve the overall survival [120]. At present, several clinical studies are ongoing on the treatment of osteosarcoma with Mifomotide, including its safety and tolerance and its efficacy in combination with other chemotherapeutic drugs (Table 2). PLX3397 (Pexidartinib) is a novel small molecule tyrosine kinase inhibitor that selectively inhibits the CSF1 receptor (CSF1R) [121]. CSF1R plays an important role in regulating the proliferation and differentiation of macrophages and is related to the prognosis of osteosarcoma [122]. Blocking CSF1R can deplete the tumor microenvironment of macrophages and enhance anti-tumor immune responses [123]. Pexidartinib has been found to be safe and tolerated in patients with advanced solid tumors [124]. Patients for a clinical trial of Pexidartinib in combination with Sirolimus for the treatment of unresectable sarcomas are currently being recruited (NCT02584647). Osteosarcoma cells express vascular cell adhesion molecule 1 (VCAM-1) [125]. The α4 integrin on the surface of TAMs can bind to VCAM-1, thereby protecting cancer cells from the effects of pro-apoptotic cytokines. Antibodies directed against α4 integrin can block this effect and prevent the growth of tumor cells in the bone marrow [126]. Natalizumab is an anti-α4 integrin antibody. The safety and tolerability of Natalizumab in children, adolescents and young adults with metastatic osteosarcoma of the lung are currently being evaluated, and positive clinical responses and overall survival rates associated with the treatment have been observed (NCT03811886).

Table 2.

Clinical trials aimed at targeting osteosarcoma macrophages and immune checkpoints

NCT Number	Compound	Target	Partner	Phase	Enrollment	Age	Status	Start Date	Completion date
NCT02441309	Mifamurtide	Macrophage	Ifosfamide	Phase 2	8	6 Years and older	Terminated	October 2014	November 4, 2016
NCT00631631	Mifamurtide (L-MTP-PE)	Macrophage	NA	NA	205	2 Years to 50 Years	Completed	January 2008	October 2012
NCT01459484	Mifamurtide	Macrophage	DDP, MTX, ADM, IFO	Phase 2	225	Up to 40 Years	Active, not recruiting	June 23, 2011	March 31, 2021
NCT02584647	PLX3397	Macrophage	NA	Phase 1	43	18 Years and older	Recruiting	November 4, 2015	March 2024
NCT03811886	Natalizumab	Macrophage	NA	Phase 1/2	20	5 Years to 30 Years	Not yet recruiting	February 2021	December 1, 2022
NCT03676985	ZKAB001	Anti-PD-L1	NA	Phase 1/2	15	18 Years to 55 Years	Recruiting	October 10, 2018	June 23, 2023
NCT04359550	ZKAB001	Anti-PD-L1	NA	Phase 3	362	12 Years and older	Not yet recruiting	June 1, 2020	June 1, 2023
NCT04668300	Durvalumab	Anti-PD-L1	Oleclumab	Phase 2	75	12 Years and older	Recruiting	December 3, 2020	June 30, 2024
NCT03006848	Avelumab	Anti-PD-L1	NA	Phase 2	15	12 Years to 49 Years	Active, not recruiting	February 16, 2017	January 31, 2023
NCT02301039	Pembrolizumab	Anti-PD-1	NA	Phase 2	144	12 Years and older	Completed	March 2015	July 1, 2020
NCT03359018	SHR-1210	Anti-PD-1	Apatinib	Phase 2	43	11 Years and older	Completed	January 1, 2018	January 30, 2020
NCT03190174	Nivolumab	Anti-PD-1	Rapamycin	Phase 1/2	40	12 Years and older	Recruiting	August 24, 2017	April 1, 2021
NCT04544995	Dostarlimab	Anti-PD-1	Niraparib	Phase 1	116	6 Months to 17 Years	Recruiting	October 6, 2020	July 23, 2030
NCT02304458	Nivolumab+ Ipilimumab	Anti-PD-1+Anti-CTLA4	NA	Phase 1/2	352	12 Months to 30 Years	Active, not recruiting	February 2, 2015	January 31, 2021
NCT02982486	Nivolumab+ Ipilimumab	Anti-PD-1+Anti-CTLA4	NA	Phase 2	60	18 Years and older	Unknown	December 2017	December 2020
NCT02500797	Nivolumab+ Ipilimumab	Anti-PD-1+Anti-CTLA4	NA	Phase 2	164	18 Years and older	Active, not recruiting	July 30, 2015	April 1, 2019

In addition, immune checkpoint inhibitors (ICIs) may affect the treatment of osteosarcoma through macrophages. The transmembrane protein CD47, expressed by osteosarcoma cells, is an innate immune checkpoint that evades the phagocytosis of TAMs by binding to the inhibitory receptor signal regulator protein α (SIRPα) on the surface of TAMs [127] (Fig. 2). Thus, an anti-CD47 monoclonal antibody may inhibit the interaction between CD47 and SIRPα to activate the phagocytosis of TAMs. Unfortunately, there are currently no clinical trials ongoing in osteosarcoma using this concept. PD-1 and CTLA-4 are expressed not only on lymphocytes, but also on monocytes and macrophages [55, 128]. It has been shown that macrophages in osteosarcoma express PD-1, and that anti-PD-1 antibodies can increase the infiltration of M1-type macrophages. By doing so, the number of osteosarcoma lung metastases can be reduced, tumor apoptosis can be enhanced and tumor cell proliferation can be inhibited [129]. A previous clinical trial of a PD-1 inhibitor (Pembrolizumab) for osteosarcoma showed that only 1 out of 22 patients responded favorably (NCT02301039). Another recently finished clinical trial of a PD-1 inhibitor for advanced osteosarcoma showed that Apatinib combined with Camrelizumab improved the PFS compared with Apatinib alone (NCT03359018). Although it did not fully achieve the intended goal, it at least suggested that a combination of multiple drugs may improve the treatment of osteosarcoma. Currently, most of the ICIs used for osteosarcoma are directed against PD-1, PD-L1 and CTLA-4. Due to the limitations of single ICIs, most clinical trials choose a combination of multiple ICIs. Among them, the combination of anti-PD-1 and anti-CTLA-4 is most common and expected to yield a better efficacy (Table 2).

Targeting immune cells in cancers considered as immune deserts

In recent years, progress in tumor immunotherapy has been closely related to the identification of immune cells in the tumor microenvironment. According to these immune cells, tumors can be classified into three phenotypes: inflamed phenotype, immune-excluded phenotype and immune-desert phenotype [130]. In the immune-desert phenotype, immune cells are distantly located from the tumor cells and, in addition, are scarce, as in osteosarcoma, neuroblastoma and other pediatric tumors [131, 132]. Therefore, the immunotherapy of these pediatric tumors significantly differs from that of adult tumors.

Current approaches aimed at targeting immune cells can be divided into: synthetic immunotherapies and therapies enhancing natural immune responses. Synthetic immunotherapies initiate new immune responses against tumor-associated targets through e.g. the use of monoclonal antibodies (mAbs) or chimeric antigen receptors (CARs). Agents that enhance natural immune responses include ICIs [133]. The use of ICIs has been found to be advantageous in adult tumors, such as non-small cell lung cancer, urothelial carcinoma and melanoma [134–136]. In these tumors, the therapeutic response rates of ICIs ranged from 15% to 45%, whereas they were not significantly efficacious in pediatric malignancies [133]. Surprisingly, it has been found that neoadjuvant chemotherapy can reprogram the tumor microenvironment to increase the degree of immune cell infiltrates in osteosarcoma, which suggests that immune-desert phenotype tumors can undergo transformation [137]. Therefore, despite the poor efficacy of single ICIs, combination therapy may be a promising approach for pediatric malignancies. Phase I clinical trials of Dostarlimab (ICI) in combination with Niraparib for osteosarcoma patients aged 6 months to 17 years have recently begun to recruit volunteers (NCT04544995). ICIs targeting pediatric tumors such as TIM-3, TIGIT and BTLA have rarely been linked to clinical trials. In pediatric malignancies, the first clinically beneficial immunotherapeutic agent was Dinutuximab, a monoclonal antibody targeting GD2 [138]. Bispecific monoclonal antibodies can simultaneously bind tumor antigens and activate T cells by binding to CD3, thereby killing the tumor cells. The bispecific monoclonal antibody Blinatumomab has, for example, shown significant efficacy in some pediatric tumors [139, 140], which offers perspectives for immunotherapy of pediatric malignancies. Dinutuximab has been shown to effectively control osteosarcoma cells and xenografts [141], which provides a basis for the design of GD2 clinical trials. It is currently in phase 1 and phase 2 clinical trials to evaluate its therapeutic efficacy in osteosarcoma and neuroblastoma (NCT02173093).

Conclusions and perspectives

Since Paget proposed the “seed and soil” theory of metastasis [74], researchers have increasingly realized that the development of tumors and the appearance of distant metastases are not only caused by the “seed” itself, but are also closely related to the “soil” in which the “seed” can survive. In addition to the role of tumor cells themselves, the progression of osteosarcoma largely depends on the surrounding TME. TAMs are among the main components of the TME, and the percentage of TAMs in osteosarcoma may even exceed 50% of the immune cell content within the TME [49], which likely accounts for its involvement in the progression of osteosarcoma. Here, we discussed data showing that TAMs may participate in the malignant progression of osteosarcoma through self-polarization, the promotion of blood vessel and lymphatic vessel formation, immunosuppression, drug resistance and other mechanisms (Fig. 3). A high expression of EGFR and other factors that promote M2 phenotype polarization of TAMs may promote the growth of osteosarcoma [67]. M2-type TAMs provide a necessary “soil” for osteosarcoma cells, thus facilitating their metastasis [78–80]. In addition, M2-polarized TAMs can activate CSCs and, thereby, promote osteosarcoma formation [65]. In the process of angiogenesis and lymphangiogenesis of osteosarcoma, IL-34 promotes M2 phenotype polarization of TAMs, thereby facilitating the formation of osteosarcoma-associated blood vessels. M2-type TAM medium treated with Resveratrol can promote the in vivo formation of lymphatic vessels in mice [99]. In addition, M2-type TAMs can induce immunosuppression by inhibiting T cell function [105]. TAMs may also secrete IL-1β, leading to drug resistance of osteosarcoma cells [112]. In terms of tumor inhibition, M1-type TAMs can inhibit the growth of osteosarcoma cells through the action of LPS + IFN-γ, while M2-type TAMs can exert antibody-dependent antitumor activity through the action of Cetuximab [69]. In addition, TAM density in the TME of osteosarcoma patients has been found to be associated with a good prognosis [117].

Fig. 3.

The role of TAMs in the progression of osteosarcoma. TAMs in osteosarcoma usually manifest themselves as tumor-promoting M2-type TAMs, which participate in the malignant progression of osteosarcoma through self-polarization, blood vessel and lymphatic vessel formation, immunosuppression and drug resistance, among others. TAMs may also show anti-tumor effects and inhibit the progression of osteosarcoma under certain conditions

However, current research on TAMs in osteosarcoma is still scarce, and the role of TAMs in osteosarcoma remains controversial. During initial stages, M1-type TAMs play a role in tumor inhibition [41, 142, 143] and M2-type TAMs play a role in tumor promotion [62, 144, 145]. However, during osteosarcoma progression, M2-type TAMs may exhibit both pro-tumor and anti-tumor properties. This duality of TAMs also exists in other tumors. In colon cancer, for example, M2-type TAMs have been found to exert anti-tumor effects via Shp2 [146], while others found that M2-type TAMs can promote colon cancer metastasis [147]. The duality of TAMs may result from the influence of other types of cells present within the TME. Clearly, TAMs cannot function alone, and parenchymal cells in the TME may concomitantly affect tumor development [148]. It has, for example, been found that active osteoclasts present at the primary site are positively correlated with the aggressiveness of osteosarcoma [149]. In addition, it has been reported that subcutaneous injection of mesenchymal stem cells lacking p53 alone or both p53 and RB in mice can cause leiomyosarcoma development, and that inoculation of the same type of cells in the bone or the periosteum can cause metastatic osteoblastic osteosarcomas [150]. Therefore, in addition to the role of TAMs in osteosarcoma progression, the question as to what extent other immune cells and/or stromal cells within the TME exert regulatory effects on TAMs, is worthy of further evaluation.

In recent years, several ideas have emerged for targeting TAMs in the treatment of various tumors. Here, we have summarized the various roles of TAMs in the progression of osteosarcoma and provided a theoretical basis for targeting TAMs to osteosarcoma. Currently, there are four main strategies for targeting TAMs: (1) inhibiting the recruitment of macrophages to tumor tissues, (2) transforming TAMs from a M2-like tumor-promoting phenotype to a M1-like anti-tumor phenotype, (3) directly killing TAMs and (4) enhancing TAM-mediated drug delivery [104]. A short summary of drugs targeting macrophages and immune checkpoints is provided in Table 2. Clinical trials of Mifamurtide are classic and numerous. In some European countries Mifamurtide, which regulates the functions of monocytes and macrophages, has been applied clinically and yielded improved overall survival rates in patients with osteosarcoma [70]. However, most treatment strategies are still in preclinical stages and have not yet evolved into mature treatment regimens. Therefore, the role of TAMs in osteosarcoma warrants further investigation. Particularly, more attention should be paid to the duality of TAMs. Due to their contrasting effects, interventions may aggravate original clinical conditions. This review has also elaborated on some other noteworthy points. In addition to the conventional polarization of M1 and M2 TAMs in osteosarcoma there are, for example, also TAMs with both phenotypes. TAMs promote the growth of blood vessels in osteosarcoma, but do not affect its prognosis. An in-depth evaluation of these different aspects may further clarify specific mechanisms underlying the diverse actions of TAMs in osteosarcoma and provide novel clues for the development of efficacious targeted treatment regimens.

Authors’ contributions

Wei Guo perceived the article, Qingshan Huang, Xin Liang, Tingting Ren, Yi Huang, Hongliang Zhang, Yuyang Yi, Chenlong Chen, Wei Wang, Jianfang Niu and Jingbing Lou performed the literature search and data analysis, and Wei Guo, Qingshan Huang and Xin Liang drafted and critically revised the work.

Declarations

Competing interest

The authors declare that they have no competing interests.