Endostatin inhibits the growth and migration of 4T1 mouse breast cancer cells by skewing macrophage polarity toward the M1 phenotype

Abstract

The phenotypic diversity of tumor-associated macrophages (TAMs) increases with tumor development. One of the hallmarks of malignancy is the polarization of TAMs from a pro-immune (M1) phenotype to an immunosuppressive (M2) phenotype. However, the molecular basis of this process is still unclear. Endostatin is a powerful inhibitor of angiogenesis capable of suppressing tumor growth and metastasis. Here, we demonstrate that endostatin induces RAW264.7 cell polarization toward the M1 phenotype in vitro. Endostatin has no effect on TAM numbers in vivo, but results in an increased proportion of F4/80⁺Nos2⁺ cells and a decreased proportion of F4/80⁺CD206⁺ cells. Overexpression of endostatin in RAW264.7 cells resulted in a decrease in the phosphorylation of STAT3, an increase in expression of vascular endothelial growth factor A and placental growth factor, and an increase in the phosphorylation of STAT1, IκBα and p65 proteins compared with controls. These results indicate that endostatin regulates macrophage polarization, promoting the M1 phenotype by targeting NF-κB and STAT signaling.

Introduction

The macrophage is a type of phagocytic cell in the immune system that is found in almost all tissues. Macrophages are derived from bone marrow progenitor cells, which mature in the peripheral blood to mononuclear cells and then migrate to various tissues and adapt to different tissue microenvironments. These mononuclear cells can further differentiate into macrophages with various phenotypes [1]. These polarization states are broadly categorized as either classically activated (M1), a pro-inflammatory and microbicidal functional phenotype induced by factors such as lipopolysaccharide (LPS) or IFN-γ [2], or alternatively activated (M2), an anti-inflammatory phenotype induced by IL-4 and IL-13, which has important functional roles in the resolution of inflammation and tissue remodeling [3].

Solid tumors are composed of malignant tumor cells and a large number of non-malignant stroma cells including endothelial cells, fibroblasts and multiple types of bone marrow cells. Most tumors have a considerable infiltration of bone marrow cells, and their activation plays a key role in tumor development [4]. Some studies demonstrated a close relationship between high numbers of tumor-associated macrophages (TAMs) and poor prognosis [4–6]. Therefore, an increasing number of studies have focused on TAMs. In a tumor, TAMs coexist with malignant cells to assist and promote tumor development and metastasis. TAMs are clustered in ischemic areas, which are necrotic/hypoxic areas of tumors, and function in the removal of necrotic cell debris [7]. TAMs respond to the hypoxic environment by altering their gene expression and exhibit a kind of pro-tumor phenotype (M2). However, studies have found that TAMs can be reprogrammed and undergo re-polarization to an anti-tumor phenotype (M1) [8–10]. Therefore, TAMs provide an ideal treatment target for preventing tumor growth and metastasis.

Endostatin, the 20-kD C-terminal fragment of collagen XVIII, is a potent inhibitor of angiogenesis through its ability to suppress endothelial cell proliferation [11]. However, several studies have shown that endostatin may also affect immune cells either directly or indirectly [12–14]. Coutinho et al. [15] reported that endostatin gene therapy of mouse tumors can induce infiltration of a large number of leukocytes. Other studies also found that endostatin treatment can increase infiltration of cytotoxic lymphocytes, such as natural killer cells and CD8⁺ cells, indicating that endostatin promotes T cell activity [13, 14].

Although studies have shown that endostatin exhibits an immunoregulation effect [11–15], whether this protein has an impact on macrophage polarization has remained unclear. In this study, we examine the biological effects of overexpressed endostatin levels on macrophages. Our results show that endostatin can induce RAW264.7 cell polarization toward an M1 phenotype by targeting the NF-κB and STAT signaling pathways.

Materials and methods

Cell lines, animals, plasmids and bacterial strains

The 4T1 mouse breast cancer cell line and the RAW264.7 mouse leukemic macrophage cell line were purchased from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences (Shanghai, China). Bone marrow-derived macrophages (BMDMs) were isolated from mouse femur and tibia as described previously [16]. Male BALB/c mice, weight approximately 18 g, were obtained from the Animal Center of Norman Bethune Medical College, Jilin University (Changchun, China). All animals were maintained under pathogen-free conditions and in accordance with protocols approved by the National Institute of Health Guide for Care. The expression vectors pcDNA3.1-ssEndostatin (pEndostatin) and pcDNA3.1-Scramble (pScramble; encoding a scrambled expression sequence used as a negative control) have been described previously [14]. The attenuated Salmonella Typhimurium (S. Typhimurium) phoP/phoQ-null bacterial strain LH430 was kindly provided by Dr. E. L. Hohmann [17]. Each plasmid was electroporated into bacteria to generate the bacterial strains SL/pScramble and SL/pEndostatin [14].

Co-culture assay for 4T1 and RAW264.7 cells

The 4T1 mouse breast cancer cells were co-cultured with RAW264.7 macrophages using a transwell culture assay. RAW264.7 cells were cultured in the upper well, and the 4T1 cells were cultured in the lower well of the chamber. After 48 h, 4T1 cells were harvested from the co-culture and re-seeded onto plates for subsequent study.

In vivo mouse tumor model of breast cancer

For the in vivo mouse tumor model, 1 × 10⁶ 4T1 cells suspended in 100 μl of sterile phosphate-buffered saline (PBS) were injected into the right flank of each mouse. Tumor size was measured from the first day until day 30 after cell injection using calipers. Tumor volume was calculated using the following formula: V (volume) = LW ² × 0.52, where “L” represents the greatest length and “W” represents the perpendicular width. When tumors reached a mean size of 250 mm³, animals were randomly split into one of four treatment groups (eight mice per group): (1) intraperitoneal injection of normal saline (control group); (2) intraperitoneal injection of attenuated S. Typhimurium phoP/phoQ (PQ group); (3) intraperitoneal injection of SL/pScramble (SL/pScramble group); and (4) intraperitoneal injection of SL/pEndostatin (SL/pEndostatin group). SL/pEndostatin and SL/pScramble bacteria were delivered to animals by two intraperitoneal injections of 5 × 10⁶ CFU/100 μl every time.

Flow cytometry (FCM) analysis

Tumor cells were washed with PBS, passed through a cell sieve and incubated with F4/80-APC, Nos2-PE and CD206-FITC (eBioscience, CA, USA) fluorescently conjugated antibodies or their respective isotype controls at 4 °C for 1 h. Co-cultured 4T1 cells were washed three times with PBS and stained with annexin-V and propidium iodide to detect apoptosis and cell cycle stage, respectively. Cells were analyzed using an Epics-XL-MCL flow cytometer (Beckman Coulter).

Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from RAW264.7 cells and tissues using Trizol reagent (Invitrogen, Carlsbad, CA, USA), and 5 µg of total RNA from each sample was used for the synthesis of complementary DNA (cDNA) using an RT-PCR kit (Promega, Madison, WI, USA). qRT-PCR was performed using SYBR Premix Ex Taq™ II (Takara, Tokyo, Japan) according to the manufacturer’s protocol. PCR reactions were run on an Applied Biosystems 7300 Real-Time PCR System (Life Technologies, MA, USA).

Enzyme-linked immunosorbent assay (ELISA)

Tumor and cell supernatants were collected, and levels of endostatin and IL-12p40 were determined using Mouse Endostatin (R&D Systems, MN, USA) and Mouse IL-12p40 (Excell Biology, Shanghai, China) ELISA kits, respectively, according to the manufacturer’s instructions.

Western blot analysis

Cells and tumors were harvested and lysed with lysis buffer (Takara). After centrifugation at 6000×g for 20 min, the protein concentration of supernatants was determined using Bradford reagent (Bio-Rad, Hercules, CA, USA). Protein samples (30 μg) were separated by SDS-PAGE and transferred onto PVDF membranes as described previously [18]. Antibodies against the following proteins were used for western blotting: VEGF (Cell Signaling, BSN, USA), Bcl-2, Bax (Calbiochem, EMD Biosciences, CA, USA), MMP-2, MMP-9 (Dako Biotech Inc., Glostrup, Denmark), STAT3, p-STAT3, STAT1, p-STAT1 and p-p65 (Abcam, JHY, UK). Membranes were incubated first with primary antibodies at 4 °C overnight, followed by HRP-conjugated secondary antibody for 1 h at room temperature. Immunoreactivity was detected using an enhanced chemiluminescence detection kit (GE Healthcare, Waukesha, WI, USA).

Immunohistochemistry (IHC) and scanning electron microscopy (SEM)

Tumors were excised from mice and processed for hematoxylin and eosin staining. Immunohistochemical analysis was carried out as described previously [18, 19] using a monoclonal antibody against F4/80 (Santa Cruz, CA, USA). SEM was performed using an S-3400N scanning electron microscope to image tumor vasculature (Hitachi Ltd., Japan).

Colony-forming cell (CFC) assay

We analyzed cell proliferation and differentiation by evaluating the ability of cells to form colonies in a semisolid medium. After co-culture, cells were washed, trypsinized and counted, and an equal cell number were then re-plated in duplicate in fresh medium in 6-well plates. Colonies of viable cells were stained with crystal violet 10 days later.

Transwell assay

A transwell system comprising a polycarbonate membrane filter with a diameter of 6.5 mm and a pore size of 8 μm (Corning Costar, Corning, NY, USA) was used to assess cell invasion as described previously [20].

MTT assays

Cells were seeded into 96-well flat-bottom microplates and cultured for various times. MTT (5 mg/ml) was added to each well, and the cells were incubated at 37 °C for 4 h. The medium was removed, and 200 μl of dimethylsulfoxide (DMSO) was added to each well. The plates were read in an ELISA reader at 490 nm, and the absorbance measurement was used to calculate cell viability.

Scratch-wound assay

A scratch wound was made by scraping a monolayer of cells cultured on a coverslip with a sterile cell lifter (3008; Corning, NY, USA). The culture medium was replaced immediately after wounding to prevent the medium from being conditioned with cell debris and factors released from detached cells. Cell migration was evaluated by measuring the size of the scratch wound at 24 and 48 h post-wounding.

Luciferase reporter assay

Cell lysates were prepared 24 h after transfection of RAW264.7 cells with the pEndostatin plasmid. Luciferase activity was subsequently detected with a Dual-Luciferase Reporter Assay Kit in accordance with the manufacturer’s instructions (Promega, Madison, WI, USA).

Statistical analyses

Data are expressed as the mean ± standard deviation (SD). Statistical comparisons and the evaluation of the significance of differences between data sets were performed using analysis of variance. All tests were two-sided. A value of p < 0.05 was considered statistically significant.

Results

Overexpression of endostatin in RAW264.7 and BMDM cells alters cell morphology and inhibits cell growth

BMDMs isolated from mouse femur and tibia were cultured for 7 days, and cell surface expression of F4/80 was determined by flow cytometry (Fig. 1a). A highly pure population of mature macrophages was observed by day 7 of in vitro culture. RAW264.7 and BMDM cells were transfected with pEndostatin plasmid, and endostatin mRNA and protein expression were examined by RT-PCR and ELISA, respectively. As expected, endostatin mRNA and protein expression levels were significantly higher in pEndostatin-transfected cells compared with the pScramble-transfected or non-transfection control (Fig. 1b, c).

Fig. 1.

Overexpression of endostatin in RAW264.7 and BMDM cells alters cell morphology and inhibits cell growth in vitro. a Analysis of BMDM purity by flow cytometry. RAW264.7 and BMDM cells were transfected with pEndostatin plasmid and expression of endostatin mRNA and protein levels were examined by b RT-PCR and c ELISA, respectively. d Light microscopy images of RAW264.7 cells transfected with pEndostatin or pScramble plasmid. e Cell viability of RAW264.7 cells transfected with pEndostatin or control plasmid as determined by MTT assay. Data represent mean ± SD from three independent experiments. *p < 0.05 versus control group; ^# p < 0.05 versus pScramble group

We examined the effects of endostatin overexpression on the morphology of transfected cells. Compared with controls, the cells of the pEndostatin group were more elongated in shape and frequently exhibited fine peripheral projections (Fig. 1d). MTT assays showed that endostatin overexpression inhibited the growth of RAW264.7 cells, with cell survival decreasing in a time-dependent manner (Fig. 1e). These findings demonstrate that endostatin alters the morphology and inhibits the growth of RAW264.7 cells.

Endostatin promotes macrophage polarization toward an M1-like phenotype

To explore the role of endostatin in macrophage polarization, we next determined the expression of M1- and M2-related polarity markers in mouse BMDM and RAW264.7 cells transfected with the pEndostatin plasmid. Analyses by qRT-PCR, ELISA and FCM showed that expression levels of the M1 markers Nos2, IL-12p40 and IL-6 were remarkably increased in pEndostatin-transfected cells compared with the control. However, the M2 markers CD206, Arg1 and IL-10 were either unchanged or exhibited a slight (but not significant) decrease in expression (Fig. 2). These observations suggest that endostatin can promote macrophage polarization toward the M1-like phenotype.

Fig. 2.

Endostatin promotes macrophage polarization in RAW264.7 and BMDM cells toward an M1 phenotype in vitro. Detection of M1- and M2-related cell surface marker expression on a RAW264.7 cells and b BMDMs by flow cytometry. Protein and RNA expression was examined by (c IL-12p40) ELISA and (d RAW264.7, e BMDM) qRT-PCR, respectively. Data represent the mean ± SD from three independent experiments. *p < 0.05 versus control group, **p < 0.01 versus control group; ^# p < 0.05 versus pScramble group

Endostatin-induced M1-like macrophages inhibit the growth, invasion and migration of 4T1 breast cancer cells in vitro

To examine the effects of M1-like RAW264.7 macrophages (induced by endostatin overexpression) on cancer cell function in vitro, a transwell culture assay was established that enabled the co-culture but physical separation of macrophages and tumor cells. To determine whether the M1-like macrophages affected the growth, invasion and migration of 4T1 cells, we used MTT, colony-forming, scratch-wound and transwell assays. The results showed that M1-like macrophages had an inhibitory effect on both tumor cell growth and migration (Fig. 3a–c). The above findings were supported by western blot data of expression of proliferation (PCNA) and invasion-related (MMPs) proteins. The 4T1 cells co-cultured with M1 macrophages exhibited a significant decrease in PCNA, MMP-2 and MMP-9 protein expression compared with controls (Fig. 3d, e).

Fig. 3.

Endostatin-induced M1-like macrophages inhibit the growth, invasion and migration of co-cultured 4T1 breast cancer cells in vitro. a Survival and colony-forming rates (left and right, respectively) of 4T1 cells. b Scratch-wound assays evaluated the migration potential of 4T1 cells following co-culture with control or endostatin overexpressing macrophages. c Transwell assays assessed the invasion potential of 4T1 cells following co-culture with control or endostatin overexpressing macrophages. d, e Western blot analysis of PCNA, MMP-2 and MMP-9 protein levels in 4T1 cells following co-culture experiments. Flow cytometry analysis of f, g cell cycle stage and h apoptosis in 4T1 cells following co-culture with control and endostatin overexpressing macrophages. Control group represents untreated RAW264.7 cells and M1 group represents RAW264.7 macrophages transfected with pEndostatin. Data represent mean ± SD from three independent experiments. *p < 0.05 versus control group; ** p < 0.01 versus control group

To elucidate the possible mechanism underlying 4T1 growth regression following co-culture with M1-like RAW264.7 macrophages, we examined cell cycle stage and apoptosis in the tumor cell population by FCM. The 4T1 cells co-cultured with M1-like macrophages were predominantly in G2 phase, whereas this population was relatively small in 4T1 cells co-cultured with control RAW264.7 cells (Fig. 3f, g). Quantitative analysis of FCM data revealed that the level of apoptosis was greater in 4T1 cells co-cultured with M1-like RAW264.7 macrophages than in 4T1 cells co-cultured with control macrophages, although this difference did not achieve significance (Fig. 3h). Taken together, these results suggest that M1-like RAW264.7 macrophages transfected with pEndostatin exert their growth suppressive effects by inducing cell cycle arrest in the tumor cell population.

Anti-tumor effect of endostatin therapy in vivo

To evaluate the effects of endostatin on tumor growth in vivo, we established a BALB/c mouse breast cancer model using the 4T1 mouse breast cancer cell line. We performed peritoneal injections of attenuated T. Salmonella harboring an endostatin expression plasmid to deliver recombinant endostatin to tumors in vivo as described in “Materials and methods” section. To confirm that this method effectively delivered endostatin to tumors in vivo, we first assessed endostatin levels within the tumors of mice. As shown in Fig. 4c, tumors isolated from mice treated with SL/pEndostatin contained high levels of endostatin compared with control groups.

Fig. 4.

Inhibition of tumor growth in vivo following endostatin treatment. a, b Images of mice after killing at day 15. c Endostatin protein expression in tumors as determined by ELISA. d The tumor index for each treatment group as measured on the final day of in vivo experimentation. e Tumor growth index values for each treatment group recorded over the 15-day study period. Tumors were measured on every third day. Data are mean ± SD from three independent experiments. *p < 0.05 versus control group; ^# p < 0.05 versus PQ group, ^Δp < 0.05 versus SL/pScramble group

The treatment of tumor-bearing mice with SL/pEndostatin had a significant suppressive effect on tumor growth compared with tumors from animals that were inoculated with salmonella harboring a control plasmid (Fig. 4a, b). The tumor growth index (which provides a measure of the rate of tumor growth) and the tumor index (which provides a measure of final tumor size on the day of sacrifice) were significantly diminished in the SL/pEndostatin treatment group compared with controls (Fig. 4d, e).

Endostatin inhibits tumor growth in vivo by skewing TAM polarization

To determine the function of endostatin in transforming the M2–M1 phenotype of TAM in vivo, tumors and peritoneal macrophages were harvested on day 15 post-treatment and expressions of the M1 markers Nos2 and IL-12p40 and the M2 marker CD206 were determined by FCM and ELISA (Fig. 5a–d). The numbers of CD206-positive and Nos2-positive peritoneal macrophages were similar among all treatment groups (Fig. 5a). Within the tumor mass, the number of TAMs (as determined by F4/80 positivity) was also unchanged among the treatment groups (Fig. 5b). However, a significant decrease in the number of F4/80⁺CD206⁺ TAMs and a significant increase in the number of F4/80⁺Nos2⁺ TAMs were observed after treatment with SL/pEndostatin (Fig. 5c). ELISA revealed a significant increase in the level of secreted IL-12p40 in the SL/pEndostatin treatment group compared with controls (Fig. 5d), a result that was consistent with the flow cytometry data. These data suggest that endostatin skews TAM polarization away from the M2-like phenotype toward the M1-like phenotype.

Fig. 5.

Endostatin inhibits tumor growth in vivo by skewing TAM polarization in vivo. Analysis of M1 and M2 markers in a peritoneal macrophages and b–d TAMs by flow cytometry and ELISA. e, f Detection of VEGF-A and PIGF protein expression by western blotting. g SEM images of tumor blood vessels (imaged in cross section) from control and SL/pEndostatin-treated mice. h Immunohistochemical analysis of tumor sections stained for F4/80 to identify TAMs. Data represent mean ± SD from three independent experiments. *p < 0.05 versus control group; ^# p < 0.05 versus PQ group

Several studies have shown that TAMs are often found within the vicinity of blood vessels of solid tumors [21, 22]. Because endostatin is a potent inhibitor of angiogenesis, we determined whether the impact of endostatin on TAM polarization was related to its anti-angiogenesis function. As shown in Fig. 5e–g, the expression of vascular endothelial growth factor A (VEGF-A) and placental growth factor (PIGF) was significantly decreased, and the lumens of blood vessels were more regular and smooth within the tumors of the SL/pEndostatin group. We also observed that the size of the necrotic area and the number of TAMs within its vicinity were significantly reduced within the tumors of SL/pEndostatin mice (Fig. 5h). These results suggest that the effect of endostatin on TAM polarization may be related to VEGF-A and PIGF signaling.

Endostatin skews RAW264.7 cell polarity toward the M1 phenotype through NF-κB and stat3 signaling

To elucidate the possible mechanism by which endostatin promotes RAW264.7 cell polarization, we examined the role of the canonical NF-κB pathway, which is known to promote the M1 polarization of macrophages [23, 24]. We performed a luciferase assay to determine the activity of NF-κB in RAW264.7 cells transfected with the endostatin expression plasmid (Fig. 6a). The results revealed an increase in NF-κB activation in the pEndostatin treatment group compared with the control. This finding was confirmed by western blot analysis of NF-κB-related proteins, which revealed that phosphorylations of p65 and IκBα were also significantly increased in the pEndostatin group (Fig. 6b–e).

Fig. 6.

Effect of endostatin on RAW264.7 macrophage polarity is associated with NF-κB signaling. a Detection of NF-κB activity by luciferase reporter assay. b–e Expression of NF-κB-related proteins VEGF and PIGF as determined by western blotting. Data represent mean ± SD from three independent experiments. *p < 0.05 versus control group; ^# p < 0.05 versus pScramble group

NF-κB drives the transcription of many pro-inflammatory genes in macrophages, such as IL-12p40, NOS2 and STAT1/2 [25–27]. In addition, the transcription of several tumor-promoting genes such as VEGF, IL-6 and cyclooxygenase 2 (COX2) is also regulated, in part, by NF-κB, supporting its crucial role in the activation of TAMs [28, 29]. We therefore examined the expression of VEGF, PIGF, STAT1 and STAT3 in control and pEndostatin RAW264.7 macrophages by western blotting. VEGF and PIGF expression as well as phosphorylation STAT3 were decreased, and phosphorylation of STAT1 was increased in the pEndostatin group compared with controls. No significant change was observed in the total levels of STAT1 and STAT3 (Fig. 6b, e). These results suggested that endostatin can promote M1-like macrophage polarization through NF-κB activation.

Discussion

Endostatin is considered one of the most potent inhibitors of angiogenesis by blocking endothelial cell proliferation, migration, invasion and tube formation. In addition, in some mouse tumor models, endostatin can significantly reduce the growth and metastasis of tumors with no obvious side effects [13, 14, 30]. Several studies have implicated endostatin in the tumor immune response [12–15]. Endostatin has been shown to increase numbers of cytotoxic lymphocytes, such as natural killer cells and CD8⁺ cells, and enhance the anti-tumor immune response [13, 14]. However, there have been no reports on the effects of endostatin on macrophages.

The macrophage content in a tumor mass can reach as high as 50 %. These cells play a key role in regulating tumor growth and angiogenesis. Tumors subjected to current cancer therapy treatments, such as radiotherapy, chemotherapy and anti-angiogenic agents, have shown a certain degree of tolerance. One possible reason is that these treatments increase the infiltration of macrophages, so that the balance of pro- and anti-angiogenic signaling is broken in the tumor environment [31]. Macrophages are classified into different phenotypes according to the Th1/Th2-type immune response. M1-like macrophages promote the inflammatory response and induce Th1 responses. These macrophages are characterized by IL-12^highIL-23^highIL-10^low expression, efficient production of effector molecules (e.g., reactive oxygen intermediates) and inflammatory cytokines (e.g., IL-1β and IL-6), and exerting cytotoxicity to phagocytic microbial pathogens and tumor cells [32]. Therefore, M1-like macrophages are considered as the soldiers of the immune system by activating the immune response to protect the host from infection and tumor development. In contrast, M2-like macrophages are a class of immunosuppressive cells that release cytokines to enhance the Th2 response. The characteristics of M2-like macrophages include Il-12^lowIL-23^lowIL-10^high expression and secretion of large amounts of CCL17, CCL22, transforming growth factor-β (TGF-β), high expression of mannose receptor and poor antigen-presenting ability [33]. As a consequence, M2 cells have a function that is more akin to a laborer, regulating inflammatory responses, adaptive immunity, angiogenesis, tissue remodeling and repair, and scavenging debris [9].

The TAM phenotype is known to diversify with tumor development. In chronic inflammation or degenerative and non-progressing tumors, M1 cells are usually the dominant phenotype. However, when the tumor is malignant, these cells acquire an M2-like phenotype [31]. In invasive tumors, TAMs cannot usually produce nitric oxide, reactive oxygen intermediates or other inflammatory cytokines (e.g., IL-12, IL-1b and Il-6). Therefore, at this time, the TAM phenotype is similar to that of M2 macrophages [34–36]. There is evidence indicating that TAMs are concentrated in the absence of vascular regions and necrotic/hypoxic regions, and their main function is to remove necrotic cell debris in these regions [7]. In addition, a study of human tumors found a positive correlation between the vessel density and the number of TAMs in the vascular region [37, 38]. There is accumulating evidence suggesting that TAM depletion can reduce the formation of tumor blood vessels [39, 40]. For example, in the PyMT mammary cancer model, genetic-scavenging macrophages can delay angiogenesis, whereas recovery of macrophage infiltration can restore blood vessel formation [40]. A genetic study showed that restoration of the expression of VEGF-A can increase the formation of tumor blood vessels in the PyMT mammary tumor model of scavenging macrophages, indicating that TAM-derived VEGF-A is essential for angiogenesis [41]. PIGF, which can trigger abnormal angiogenesis, inhibits T cell-mediated tumor immunity and skews TAM polarity toward a pro-tumor and pro-angiogenic M2-like phenotype [42–44]. A study by Peng [49] found that endostatin can change the polarization of TAMs by promoting the normalization of tumor blood vessels. These findings indicate that VEGF-A and PIGF are key regulators of pro-angiogenic TAM activity. In this study, we found that endostatin induced RAW264.7 cell and BMDM polarization toward the M1-like phenotype in vitro. Although endostatin had no effect on TAM numbers in vivo, an increase in the proportion of F4/80⁺Nos2⁺ cells and a decrease in the proportion of F4/80⁺CD206⁺ cells were observed. In addition, the expressions of VEGF-A and PIGF were significantly decreased in the endostatin treatment group. We propose that the endostatin-induced polarization of macrophages toward an M1 phenotype is mediated largely by endostatin-induced down-regulation of PIGF and VEGF-A.

Several studies have shown that NF-κB is an important decisive factor in the balance between the pro- and anti-tumor properties of macrophages, so NF-κB can be targeted to “re-educate” tumor-promoting macrophages to exert anti-tumor functions [45, 46]. The p65/p50 heterodimer of NF-κB is the mediator of NF-κB transcriptional activity for pro-inflammatory gene expression, and the classical NF-κB pathway drives the expression of many pro-inflammatory genes in macrophages, including IL-12p40 and STAT1 [25–27]. However, STAT3 activation is not only related to the transcription of M2 genes (e.g., Arg1 and IL-6), but can inhibit NF-κB [47, 48]. Our data suggest that endostatin inhibits the phosphorylation of STAT3 and promotes the phosphorylation of STAT1, p65 and IκBα.

In summary, our findings indicate that endostatin regulates the M1 polarization of RAW264.7 macrophages by targeting NF-κB and other genes associated with its downstream signaling pathway in vitro. Endostatin may inhibit the growth of mouse breast cancer in vivo by regulating the polarization of TAMs via two possible mechanisms: first by skewing the polarity of TAM from an M2-like to an M1-like functional phenotype; or second, by the specific inhibition of M2 polarity, thereby increasing the proportion of M1-like TAMs. Further research is required for a more thorough understanding of the mechanisms by which endostatin controls macrophage polarity in vivo.

Abbreviations

APC

Allophycocyanin

Arg1

Arginase-1

BMDM

Bone marrow-derived macrophage

CD206

Mannose receptor 1

cDNA

Complementary DNA

CFC

Colony-forming cell

COX2

Cyclooxygenase 2

DMSO

Dimethylsulfoxide

ELISA

Enzyme-linked immunosorbent assay

FBS

Fetal bovine serum

FCM

Flow cytometry

FITC

Fluorescein isothiocyanate

IHC

Immunohistochemistry

IL-6

Interleukin 6

IL-10

Interleukin 10

IL-12p40

Interleukin 12p40

M1

Classical activation macrophage

M2

Alternative activation macrophage

MMP

Matrix metalloproteinase

MTT

3-(4,5-Dimethylthiazol-2-yl)-2,5-dimethyltetrazolium bromide

NF-κB

Nuclear factor κB

NOS2

Inducible nitric oxide synthase

PBS

Phosphate-buffered saline

PE

Phycoerythrin

PI

Propidium iodide

PIGF

Placenta growth factor

qRT-PCR

Quantitative real-time reverse transcription polymerase chain reaction

SEM

Scanning electron microscopy

STAT

Signal transducer and activator of transcriptions

TAMs

Tumor-associated macrophages

VEGF

Vascular endothelial growth factor

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.