Coagulation Factor X Regulated by CASC2c Recruited Macrophages and Induced M2 Polarization in Glioblastoma Multiforme

Tumor-associated macrophages (TAMs) constitute a major component of inflammatory cells in the glioblastoma multiforme (GBM) tumor microenvironment. TAMs have been implicated in GBM angiogenesis, invasion, local tumor recurrence, and immunosuppression. Coagulation factor X (FX) is a vitamin K-dependent plasma protein that plays a role in the regulation of blood coagulation. In this study, we first found that FX was highly expressed and positively correlated with TAM density in human GBM. FX exhibited a potent chemotactic capacity to recruit macrophages and promoted macrophages toward M2 subtype polarization, accelerating GBM growth. FX bound to extracellular signal-related kinase (ERK)1/2 and inhibited p-ERK1/2 in GBM cells. FX was secreted in the tumor microenvironment and increased the phosphorylation and activation of ERK1/2 and AKT in macrophages, which may have been responsible for the M2 subtype macrophage polarization. Moreover, although the lncRNA CASC2c has been verified to function as a miR-101 competing endogenous RNA (ceRNA) to promote miR-101 target genes in GBM cells, we first confirmed that CASC2c did not function as a miR-338-3p ceRNA to promote FX expression, and that FX was a target gene of miR-338-3p. CASC2c interacted with and reciprocally repressed miR-338-3p. Both CASC2c and miR-388-3p bound to FX and commonly inhibited its expression and secretion. CASC2c repressed M2 subtype macrophage polarization. Taken together, our findings revealed a novel mechanism highlighting CASC2c and FX as potential therapeutic targets to improve GBM patients by altering the GBM microenvironment.

Tumor-associated macrophages (TAMs) constitute a major component of inflammatory cells in the glioblastoma multiforme (GBM) tumor microenvironment. TAMs have been implicated in GBM angiogenesis, invasion, local tumor recurrence, and immunosuppression. Coagulation factor X (FX) is a vitamin K-dependent plasma protein that plays a role in the regulation of blood coagulation. In this study, we first found that FX was highly expressed and positively correlated with TAM density in human GBM. FX exhibited a potent chemotactic capacity to recruit macrophages and promoted macrophages toward M2 subtype polarization, accelerating GBM growth. FX bound to extracellular signal-related kinase (ERK)1/2 and inhibited p-ERK1/2 in GBM cells. FX was secreted in the tumor microenvironment and increased the phosphorylation and activation of ERK1/2 and AKT in macrophages, which may have been responsible for the M2 subtype macrophage polarization. Moreover, although the lncRNA CASC2c has been verified to function as a miR-101 competing endogenous RNA (ceRNA) to promote miR-101 target genes in GBM cells, we first confirmed that CASC2c did not function as a miR-338-3p ceRNA to promote FX expression, and that FX was a target gene of miR-338-3p. CASC2c interacted with and reciprocally repressed miR-338-3p. Both CASC2c and miR-388-3p bound to FX and commonly inhibited its expression and secretion. CASC2c repressed M2 subtype macrophage polarization. Taken together, our findings revealed a novel mechanism highlighting CASC2c and FX as potential therapeutic targets to improve GBM patients by altering the GBM microenvironment.
Keywords: tumor-associated macrophages, polarization, glioblastoma multiforme, extracellular signal-related kinase 1/2, aKT inTrODUcTiOn Glioblastoma multiforme (GBM) is the most common type of primary brain tumor in adults and is associated with poor prognosis (1,2). One challenge of GBM is genetic heterogeneity, while noncancerous stromal cells in the tumor microenvironment are genetically stable as therapeutic targets (3). Tumor-associated macrophages (TAMs) in particular are associated with high tumor grades and poor prognosis in many cancers, including GBM (4). Abundant macrophages infiltrate GBM and promote tumor progression in multiple aspects. TAMs secrete cytokines, including interleukin-6 (IL-6), interleukin-10 (IL-10), tumor necrosis factor-α (TNF-α), and interferon-γ, which have been shown to promote tumor cell growth (5). TAMs could also facilitate angiogenesis by releasing vascular endothelial growth factor-α and associating with adjacent endothelial tip cells, facilitating vascular anastomosis (6). A population of TIE2 + macrophages also promotes tumor cell intravasation into circulation by alignment along the vessels (7). TAMs also secrete various cytokines and chemokines that suppress CD4 + and CD8 + T cell effector function directly or indirectly by recruiting regulatory T cells to the tumor microenvironment (8).
Coagulation factor X (FX) is a vitamin K-dependent plasma protein known to be an important player in the regulation of blood coagulation by converting prothrombin into thrombin (15). Activated FX (FXa) occupies a central position in the coagulation cascade and plays a role in tissue remodeling, fibrosis, and cancer via activating protease-activated receptors (PAR)-1 or PAR-2 to mediate intracellular signaling (16,17). Classical ly, FXa-induced PAR signaling induces phosphoinositide hydrolysis, leading to calcium oscillation. FXa also triggers the phosphorylation of mitogen-activated protein kinases (MAPKs), specifically extracellular signal-related kinase (ERK) and c-Jun N-terminal kinase, activates the PI3K-AKT/PKB pathway and the phosphorylation of mTOR, leading to cell proliferation, differentiation, and migration (18). Furthermore, FXa regulates inflammatory signaling by inducing the expression of IL-6, IL-8, monocyte chemotactic protein-1, and intracellular adhesion molecule (19). Many observations have shown ectopic expression of FX in cancer cells, including ovarian cancer, small lung cell carcinoma, renal cell carcinoma, and malignant melanoma (20). Our previous studies have indicated that FX overexpression in glioma was due to promoter hypomethylation, and its protein expression correlated with tumor grade and overall survival (21).
In this study, we demonstrated that FX had chemotactic ability that recruited macrophages in GBM and mainly promoted macrophage polarization to M2 subtype, facilitating tumor growth. Furthermore, FX interacted with ERK1/2 and decreased p-ERK1/2 in GBM cells, while it was secreted into the tumor microenvironment and increased p-ERK1/2 and p-AKT in macrophages, which played a role in macrophage polarization.

MaTerials anD MeThODs cell culture
The human astrocytoma cell line U251 and mouse glioma cell line GL261 were purchased from cell banks of the Chinese Academy of Sciences (Shanghai, China). The normal human astrocyte cell line HEB was obtained from the Guangzhou Institute of Biomedicine and Health, Chinese Academy of Sciences (Guangzhou, China) (22). Primary cultured GBM cells (G1124, G1104) (23) were separated from human GBM samples by the Department of Neurosurgery, Xiangya Hospital, Central South University. All cells were cultured in Dulbecco's modified Eagle's medium (DMEM, HyClone) supplemented with 10% fetal bovine serum (FBS, Biological Industries) and 1% penicillin/streptomycin (HyClone) at 37°C and 5% CO2 in a humidified atmosphere.

Plasmids
Factor X was amplified from G1124 cells and cloned into plasmids pEGFP-C1, p3xFLAG-CMV-10, and pcDNA3.1. ERK1 and ERK2 were cloned from 293 cells and fused into pDsRed1-N1 plasmid. The 3′UTR regions of FX and CASC2c were synthesized by Sangon Biotech Company and inserted into a pmirGLO Vector.

Transient Transfection and lentivirus infection
Transient transfection of miRNA mimics and plasmids was performed according to the manufacturer's manual using lipofectamine 3000 reagent (Thermo Fisher Scientific, L3000015). The lentivirus system purchased from Invitrogen contained four FX Recruited M2 TAMs in GBM Frontiers in Immunology | www.frontiersin.org July 2018 | Volume 9 | Article 1557 plasmids: pLVX-mCherry-N1, pLP1, pLP2, and pLP/VSVG. FX was constructed in pLVX-mCherry-N1 and transfected into 293FT cells with pLP1, pLP2, and pLP/VSVG. The cellular supernatants were harvested after 48 and 72 h and ultracentrifugation to collect the lentivirus. We infected GL261 cells with lentivirus and screened positive cells with puromycin (Sigma-Aldrich). Then, the cells were cultured in DMEM with 10% FBS (HyClone).

real-Time Pcr analysis of mirna and mrna
Total RNA was extracted from cultured cells using the TRI reagent (Molecular Research Center, MRC). Total RNA (2 µg) was reverse transcribed to cDNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) according to the manufacturer's procedure. Real-time PCR was performed using SYBR Green PCR kits (Bimake). miRNA was reverse transcribed to cDNA using a miScript reverse transcription kit (GenePharma). Expression of miRNA was measured by real-time PCR using the miRNA Real-Time PCR Assay Kit (GenePharma). The sequences of the primers are listed in Table  S1 in Supplementary Material.

co-immunoprecipitation
HEK293 cells were lysed with IP lysis buffer (10 mM Tris-HCl pH 7.5; 300 mM NaCl; 10 mM EDTA; 0.5% Triton X-100) supplemented with protease inhibitor cocktail and phosphatase inhibitor. Cell lysates were incubated with Flag antibody for 12 h at 4°C. Then, the solution was incubated with Protein G beads (Thermo Fisher Scientific) for 4 h at 4°C. After the beads were washed and boiled, the supernatants were collected for Western blot detection.

Monocytes/Macrophages chemotaxis assays
THP-1 cells and mouse monocytes/macrophages were cultured in the RPMI-1640 (HyClone) with 10% FBS (HyClone). THP-1 cells were primed with 50 nM phorbol 12-myristate 13-acetate (PMA, Sigma) for 48 h to become monocyte-derived macrophages. Chemotaxis assays assessing cell chemotactic potential were performed in 24-well plates with an 8-µm aperture. A total of 5 × 10 5 primed THP-1 cells or mouse monocytes/macrophages were cultured in the upper chamber, and culture supernatants from GBM cells were added to the lower chamber and cultured for 24 h. Cells in the lower chamber were fixed with 4% paraformaldehyde and stained with crystal violet. Photos of the cells were captured using a microscope system (Olympus).

immunofluorescence staining and confocal laser scanning
For analysis of the correlation between FX and Iba1, frozen astrocytoma tissue sections were permeated with 0.25% Triton X-100 for 10 min and blocked with 10% goat serum for 30 min. Then, the sections were stained with anti-FX (Thermo Fisher Scientific) and anti-Iba1 (Thermo Fisher Scientific) antibodies for 12 h at 4°C followed by staining with Alexa Fluor 488-and 594conjugated antibodies (Thermo Fisher Scientific, A11029, A27016). After staining with DAPI (Beyotime Biotechnology, C1002), the sections were mounted by anti-fluorescence quenching agent. Confocal analysis was performed on the Ultra-VIEW VoX system (PerkinElmer) according to the manufacturer's instructions.
immunohistochemistry Human astrocytoma and mouse orthotopic tumor paraffin sections were dewaxed, rehydrated, and subjected to antigen retrieval. Sections were blocked with 3% hydrogen peroxide for 10 min and normal goat serum for 1 h at room temperature. Then, the sections were incubated with anti-FX (Thermo Fisher Scientific), Iba1 (Thermo Fisher Scientific), CD163, and CD11c (Proteintech) antibodies for 12 h at 4°C and incubated with biotinylated secondary antibody (Maxim Biotechnologies) for 30 min at room temperature followed by streptavidin-conjugated HRP (Maxim Biotechnologies) for 30 min. Staining was visualized with 3,3′-diaminobenzidine (Maxim Biotechnologies) and counterstained with hematoxylin. The immunohistochemical scoring was performed according to the Konno's criteria (24). The staining index (0-12) of FX was determined by multiplying the score of staining intensity with score of positive area. The staining intensity was scored as 0, negative; 1, weak; 2, moderate; and 3, intense. The positive area was defined as 0, less than 5%; 1, 6-25%; 2, 26-50%; 3, 51-75%; and 4, greater than 75%. FX expression quantification was analyzed by Image Pro Plus vision 6.0, and the IOD was used for drawing. Iba1, CD11c, and CD163 expression were quantified as follows: the number of Iba1 + cells, CD11c + cells, and CD163 + cells were counted in three random images from a single section and the number per square millimeter was used for drawing.

ccK8, eDU incorporation, and Transwell assay
The CCK8 (Bimake) and EDU incorporation assays (Ribobio) were performed according to the manufacturers' procedures. Cell viability was assessed by a CCK8 assay as previously described (25). The proliferation ability of astrocytoma cells was detected by an EDU incorporation assay as previously described (26). The invasive ability of astrocytoma cells was tested by a transwell assay (Corning Inc.) as previously described (25).

intracranial implantation Mouse Model
All animal experiments were approved by the Animal Care and Use Committee of Central South University. Five-week-old female C57BL/6 mice were anesthetized with intraperitoneal sodium pentobarbital (40 mg/kg) and fixed in a stereotaxic instrument. Then, an incision was made on the midline of the mouse head, and a hole was drilled in the right hemisphere at AP = +1 mm and ML = −2.5 mm from bregma. Five microliters of 10 6 cells were injected into the brain at a depth of −3.5 mm from the dura. Mouse weight and survival were recorded daily. After the mice were sacrificed, the whole brains were fixed with 4% paraformaldehyde. HE staining and immunohistochemistry were performed according to the standard procedures.

statistical analysis
All the experiments were repeated at least three times, and the representative data are shown. The statistical analysis was performed using GraphPad Prism 5 and SPSS version 17.0. Data analysis was performed with Student's t-test and one-way ANOVA and presented as the mean ± SEM. p Values less than 0.05 were considered significant.

FX Promoted the growth of gBM cells In Vivo but Did not affect cell Proliferation In Vitro
Our previous research reported that FX was a hypomethylation gene highly expressed in glioma (21). In this study, to ascertain the potential role of FX in the pathogenesis of astrocytoma, astrocytoma tissue sections of different WHO grades were detected by immunohistochemistry with an FX antibody. FX was not detected in human normal brain tissues, while FX expression was increased significantly in astrocytoma tissues and dramatically elevated in high-grade astrocytoma (WHO III and IV grade) compared with that in low grade (WHO I and II grade) astrocytoma ( Figure 1A).
To confirm the expression of FX in cell lines, FX protein levels were measured in primary cultured GBM cells (G1124, G1104), U251, normal human glial cells (HEB), and HEK293 cells. FX was highly expressed in G1124 cells and, to a lesser extent, in G1104 cells, but it was nearly undetectable in U251, HEB, and HEK293 cells ( Figure 1B). Unlike GFP alone, which filled whole cells, FX-GFP localized primarily to the trans-Golgi network and vesicles ( Figure S1A in Supplementary Material), suggesting that FX was a secreted protein.
To examine the role of FX in GBM cells, FX-shRNAs were transfected into G1124 cells, and decreased FX expression was detected by Western blot ( Figure 1C). Knockdown of FX did not affect the proliferation of G1124 cells measured using CCK8 ( Figure 1D) and EDU assays ( Figure 1E). Matrigel invasion assays showed that the invasion of G1124 cells was not influenced by FX ( Figure 1F). Furthermore, overexpressing FX in U251 cells did not influence the cell proliferation and invasion (Figures 1G-J). Next, GL261 cells that stably overexpressed FX by lentiviral infection were constructed and transplanted into the corpus striatum zone of C57BL/6 mice to form intracranial orthotopic GL261 xenografts. The expression of FX was higher in GL261-FX cells than in GL261-CON cells ( Figure S1B in Supplementary Material). These two cell lines had the same proliferation rate despite FX expression ( Figure S1C in Supplementary Material). By contrast, overexpression of FX significantly promoted the growth of intracranial orthotopic GL261 xenografts measured by HE staining ( Figure 1K). Overexpression of FX significantly enhanced the expression of Ki-67 as shown by immunohistochemistry ( Figure 1L), which suggested that FX promoted cell proliferation in vivo. These results indicated that increasing FX expression accelerated glioma tumor growth in vivo, while FX did not affect the proliferation and invasion in vitro.

high expression of FX Was Positively correlated With TaM Density in gBM
Recent studies indicated that FX is a secretion protein that plays a key role in blood coagulation by converting prothrombin to thrombin, leading to the formation of a fibrin clot. In addition, integrin αMβ2 is a key receptor of FX in mediating cell migration (27). Integrin subunit alpha M, also known as cluster of differentiation molecule 11b (CD11b), encodes the integrin alpha M chain, which is frequently used as a marker of macrophages. Therefore, we speculated that FX may influence the chemotaxis of macrophages to facilitate tumor growth in the brain. To address the correlation between FX and TAMs in human astrocytoma, astrocytoma sections were immunostained with FX and the TAM marker Iba1. In WHO grades I and II astrocytoma, TAM and FX expression was scarce, while in WHO grades III and IV astrocytoma, TAM and FX were highly expressed (Figures 2A-C).
Immunofluorescence and immunohistochemistry assays further indicated that higher FX expression was accompanied by more TAM infiltration, while lower FX expression in human GBM tissues had less Iba1 staining (Figures 2D-F). In GL261-FXderived xenografted tumors, TAMs were more abundant than that in GL261-CON-derived tumors ( Figure 2G). These results suggested a positive correlation between FX protein levels and TAM density in GBMs, and FX may contribute to the infiltration of TAMs to promote tumor growth in vivo.

FX exhibited a Potent capacity to recruit Macrophages/Monocytes
To examine whether FX function as a potent chemoattractant of TAMs, a series of migration assays were used to examine the (g) Western blotting showed that overexpression of FX increased FX expression in U251 cells. (h-J) FX overexpression did not affect U251 cell viability, proliferation, and invasion measured by CCK8 (h), EDU incorporation assays (i), and transwell assays (J). Data are presented as mean ± SEM (*p < 0.05). (K) GL261 cells that were infected with lentivirus (GL261-CON and GL261-FX) were intracranially injected into the corpus striatum of C57BL/6 mice. HE staining showed that the mice transplanted with GL261-FX cells exhibited larger tumor volumes. Data are presented as the mean ± SEM (*p < 0.05). (l) Immunohistochemistry detected FX and Ki67 expression in intracranial xenografts.  Figures 3A,B). The culture supernatants from G1124 cells which knockdown of FX reduced the migration of PMA-primed macrophage-like THP-1 cells (Figure 3C). Preincubation of G1124 cell culture supernatants with an anti-FX antibody attenuated the chemotaxis effect of FX on macrophage migration ( Figure 3D). By contrast, overexpression of FX in U251 cells increased FX expression in the cells and culture supernatants (Figures 3E,F), and the culture supernatants markedly enhanced the ability to recruit PMA-primed macrophage-like THP-1 cells ( Figure 3G). To further explore whether the FX-mediated chemotactic effect on macrophage migration was dose dependent, we performed the chemotaxis experiments with different concentrations of recombinant FX (rFX) protein.
The migration of macrophages toward FX was significantly enhanced as the rFX protein increased (Figure 3H), indicating that FX attracted macrophages in a dose-dependent manner. The capacity of FX to recruit macrophages was further demonstrated by the isolated mouse primary macrophages/ monocytes in migration and invasion assays. Overexpressing FX in GL261 cells recruited more macrophages/monocytes to the lower chamber of the transwells ( Figure 3I). The chemotactic ability of FX was attenuated by FX antibody (Figure 3J).
The above data demonstrated that FX secreted by GBM cells exhibited a potent capacity to recruit macrophages/monocytes. To address whether FX recruited macrophages in an mTORdependent or -independent manner, we treated PMA-primed macrophage-like THP-1 cells with the culture supernatants from G1124 cells in which FX had been knocked down or culture supernatants from U251 cells in which FX was overexpressed and detected the expression of p-mTOR and p-p70S6K of THP-1 cells. When PMA-primed THP-1 cells were treated with G1124-sh-FX cells supernatants, p-mTOR and p-p70S6K decreased. By contrast, when PMA-primed THP-1 cells were incubated with U251-FX cells supernatants, p-mTOR and p-P70S6K increased. Moreover, when PMA-primed THP-1 cells were treated with recombinant protein FX, p-mTOR and p-p70S6K increased (Figures 3K-N). In addition, inhibitor of mTOR rapamycin suppressed p-mTOR and p-p70S6K in THP-1 cells. But rapamycin did not inhibit the chemotaxis of FX on PMA-primed THP-1 cells (Figures 3O,P) suggested that FX recruited macrophages in an mTOR-independent manner.

FX Overexpression in gBM cells specifically increased M2 Tumorsupportive TaMs
To address which subtype of TAMs was recruited or maintained by FX in GBM, we applied M1-specific marker (CD11c) and M2-specific marker (CD163) to distinguish TAMs in intracranial orthotopic GL261-CON-and GL261-FX-derived xenografts.

FX recruited and influenced the Polarization of Macrophages Through erK1/2 and aKT
To further investigate the molecular mechanism underlying the FX-mediated recruitment and polarization of TAMs, we used CD11b antibody to block the integrin signaling on PMA-primed macrophage-like THP-1 cells and found that CD11b antibody significantly reduced the chemotaxis and mobility of the THP-1   (Figure 5A). At the same time, CD11b antibody increased the expression of the M1 markers (CXCL9, IL-12β, and CXCL10) and decreased the expression of M2 markers (ARG1, MRC1, and STAB1) ( Figure 5B). Furthermore, ERK1/2 and AKT played important roles in the regulation of M2 macrophage cell-specific genes such as ARG-1 and MRC-1 (28). Prediction with bioinformatics software Scansite 3.0 found that FX may interact with ERK1/2 ( Figure S2A in Supplementary Material). Therefore, we detected whether FX colocalized with ERK1/2. GFP-FX and RFP-ERK1 or RFP-ERK2 were constructed and cotransfected into HEK293 cells. As shown in Figure 5C, confocal fluorescence microscopy displayed that FX was colocalized with ERK1 and ERK2. To determine whether FX and ERK1/2 could be co-immunoprecipitated from cells, we constructed a full-length Flag-FX vector and transfected in HEK293 cells. FX was co-immunoprecipitated with endogenous ERK1 and ERK2 but not p-ERK1 and p-ERK2 ( Figure 5D). In addition, FX was co-immunoprecipitated with exogenous ERK1 and ERK2 ( Figure 5E). GST with an FX fusion protein was constructed, and a GST pull-down assay demonstrated that FX bound to ERK1 and ERK2 (Figure 5F). We also used co-immunoprecipitation to detect whether AKT interacted with FX, and the results suggested that AKT did not interact with FX ( Figure S2B in Supplementary Material).
We further investigated whether FX mediate macrophage polarization through ERK1/2 and AKT. p-ERK and p-AKT were decreased when THP-1 cells were treated with cell culture supernatants from G1124 cells in which FX was knocked down, while when THP-1 cells were treated with cell culture supernatants from U251 cells which FX was overexpressed, p-ERK1/2 and p-AKT were increased ( Figure 5G). p-ERK1/2 and p-AKT also increased when THP-1 cells were treated with rFX protein in a concentration-dependent manner ( Figure 5H). Next, we used an inhibitor of p-ERK1/2 (GSK2606414) or p-AKT (MK2206) to inhibit p-ERK1/2 or p-AKT, respectively, in THP-1 cells and measured the markers of M1 and M2 macrophages. When THP-1 cell were incubated with FX-overexpression-U251 cell supernatants supplemented with these two inhibitors, both inhibitors reversed the M1 and M2 macrophage marker expression induced by FX ( Figure 5I). Subsequently, we also examined the effect of FX on intracellular p-ERK1/2 and p-AKT in GBM cells. Overexpression of FX decreased p-ERK1/2 in G1104 and U251 cells, while knockdown of FX increased p-ERK1/2 in G1124 cells but had almost no influence on p-AKT ( Figure 5J). These results demonstrated that FX bound to ERK1 and ERK2 and is involved in regulating ERK1/2 phosphorylation.
Taken together, these findings indicate that FX bound to ERK1/2 and inhibited p-ERK1/2 expression in GBM cells. Moreover, FX was secreted from GBM cells and bond to integrin αMβ2 on the surfaces of macrophages to recruit them to the tumors. FX raised expression of p-ERK1/2 and p-AKT in macrophages to induce them to M2-type polarization.

mir-338-3p Targeted FX to suppress Macrophage Migration
We confirmed that FX recruited macrophages to the tumors and influenced polarization of macrophages through p-ERK1/2 and p-AKT. Then, we turned our attention to the molecule that affected the expression of FX in GBM cells. By prediction with TargetScan, 1 FX may be targeted by miR-338-3p ( Figure S3A in Supplementary Material); moreover, according to the miRcode, 2 miR-338-3p was likely to interact with lncRNA CASC2c, which is an onco-RNA acting in tumorigenesis of astrocytoma ( Figure  S3B in Supplementary Material). miR-338-3p mimics were transfected into G1124 cells that highly expressed FX, and the expression of FX was decreased for mRNA ( Figure 6A) and protein ( Figure 6B). To ascertain whether FX was the direct target gene of miR-338-3p, FX putative miR338-3p recognition sequences (pmirGLO-FX-WT) and mutant derivatives lacking the binding sequences (pmirGLO-FX-MUT) were cloned downstream of the luciferase gene into the pmirGLO vector and were transfected into HEK293 cells with NC or miR-338-3p mimics FigUre 4 | Factor X (FX) promoted macrophage polarization to the M2 subtype. (a) Immunohistochemical staining of GL261-CON-and GL261-FX-derived xenografted tumors by CD11c and CD163 antibody. In GL261-FX-derived xenografted tumors, there was more CD163 staining and less CD11c staining than in the GL261-CON tumor. (B) Macrophage subtype markers were measured by real-time PCR after treatment with supernatants from G1124 cells transfected with sh-NC or sh-FX. In the sh-FX group, IL-1β, IL-12α, CXCL9, IL-12β, and CXCL10 levels were higher, while LYVE1, MRC1, and STAB1 levels were lower compared those in the sh-NC group (mean ± SEM, *p < 0.05, **p < 0.01). (c) Supernatants from U251 cells transfected with FX or control vectors were added to THP-1 cells, and the markers' mRNAs were detected by real-time PCR. IL-1β, CXCL9, IL-12β, and CXCL10 levels were decreased, while HMOX1, MRC1, and SerpinB2 levels were increased when FX was overexpressed (mean ± SEM, *p < 0.05, **p < 0.01). (D) THP-1 cells were treated with recombinant protein FX and markers' mRNAs were measured by real-time PCR. CXCL10 expression decreased, while M2 markers (LYVE1, MRC1, SerpinB2, and STAB1) increased after treatment with recombinant FX (rFX) (mean ± SEM, *p < 0.05, **p < 0.01).  Figure 6C). To further validate whether miR-338-3p could regulate the migration of macrophages, the supernatants from G1124 cells transfected with miR-338-3p or NC were harvested to chemoattract THP-1-derived macrophage-like cells. miR-338-3p decreased the number of THP-1 cells that migrated ( Figure 6D). However, the M1 markers (IL-1β, IL12A, CXCL9, CXCL10, and iNOS) and M2 markers (ADM, HMOX1, MRC1, SerpinB2, and STAB1) were almost unchanged in THP-1 cells, whether the cells were treated with supernatants from G1124 cells transfected with miR-338-3p or NC (Figure 6E). Flow cytometry also demonstrated that the proportion of CD11b + Cd80 + M1 macrophages and CD11b + CD206 + M2 macrophages did not changed in THP-1 cells treated with supernatants from G1124 cells transfected with miR-338-3p compared with NC (Figures 6F,G). These results suggested that miR-338-3p targeted FX and suppressed macrophage migration through FX, but did not affect macrophage polarization.
To explore CASC2c's function on macrophage migration and polarization, the supernatants from G1124 cells transfected with CASC2c were collected to treat PMA-primed THP-1 cells. CASC2c suppressed the migration of macrophages ( Figure 7F). M2 markers decreased while M1 markers increased compared with those of control cells ( Figure 7G). Conversely, we used sh-CASC2c to knockdown CASC2c mRNA expression to further verify these results. Knockdown of CASC2c increased the mRNA ( Figure 7H) and protein ( Figure 7I) expression of FX in U251 cells. The number of THP-1 cells that migrated was increased when THP-1 cells were incubated with supernatant from U251 cells with CASC2c knockdown (Figure 7J). The mRNA expression of M1 markers decreased while that of M2 markers increased ( Figure 7K). Flow cytometry also showed that the proportion of CD11b + Cd80 + M1 macrophages increased and CD11b + CD206 + M2 macrophages decreased when THP-1 treated with supernatants from G1124 cells transfected with CASC2c compared with control vector (Figures 7L,M). These results suggested that CASC2c regulated FX expression and inhibited macrophage migration and polarization to the M2 subtype.

DiscUssiOn
Glioblastoma multiforme often contains a large number of TAMs that constitute a major component of the inflammatory cells in the tumor microenvironment (29). These cells have been implicated in glioma angiogenesis, invasion, local tumor recurrence, and immunosuppression (30). Tumor cells can secrete several chemokines, such as CC chemokine ligand 2, soluble colony-stimulating factor 1, stromal cell-derived factor, hepatocyte growth factor, and periostin (POSTN), to recruit TAMs in cancers (31)(32)(33)(34). FXa is activated by the cleavage of the activation peptide by factor IXa (in the intrinsic pathway) or by factor VIIa (in the extrinsic pathway). FXa is often characterized as a hemostatic agent and induces formation of thrombin during thrombosis. Thrombin activates tumor cell adhesion to platelets and endothelial cells and subsequently enhances tumor cell metastasis and angiogenesis (35). Jiang et al. found that adding the TF-FVIIa-FXa complex or FXa alone to the lower chamber promoted MCF-7 cell migration from the upper The luciferase reporter assay showed that miR-338-3p decreased luciferase activity (mean ± SEM, **p < 0.01). (D) The number of phorbol 12-myristate 13-acetate (PMA)-primed THP-1 cells migrating toward G1124 cell supernatants decreased when G1124 cells were transfected with miR-338-3p mimics (mean ± SEM, *p < 0.01). (e) Real-time PCR detected the expression of M1 and M2 markers in THP-1 cells after incubation with G1124 cell supernatants transfected with NC or miR-338-3p mimics (mean ± SEM). (F,g) Flow cytometry analysis showed that the proportion of CD11b + CD80 + M1 macrophages and CD11b + CD206 + M2 macrophages did not changed in THP-1 cells treated with supernatants from G1124 cells transfected with miR-338-3p compared with NC. (g) Bar graph of CD11b + CD80 + and CD11b + CD206 + cell proportion in panel (F). FigUre 5 | Factor X (FX) recruited and influenced macrophage polarization through extracellular signal-related kinase (ERK)1/2 and AKT. (a) Transwell assays showed that THP-1 cells' migration toward FX decreased after preincubation with a CD11b antibody (mean ± SEM, **p < 0.01). (B) Real-time PCR analysis of M1 and M2 markers in THP-1 cells after the cells were treated with CD11b antibody (mean ± SEM, *p < 0.05, **p < 0.01). (c) Confocal fluorescence microscopy of HEK293 cells cotransfected with GFP-FX (green) and RFP-ERK1 (red) or RFP-ERK2 (red) showed the co-localization of FX with ERK1 and ERK2. (D) Coimmunoprecipitation showed the interaction between FX and endogenous ERK1/2 in HEK293 cells after transfection with Flag-FX. (e) Co-immunoprecipitation showed the interaction between exogenous FX and ERK1/2 in HEK293 cells after cotransfection with pcDNA3.1-FX and flag-ERK1 or myc-ERK2. (F) GST pull-down assays showed the binding of FX and ERK1 or ERK2. (g) p-ERK1/2 and p-AKT were decreased when THP-1 cells were treated with G1124 cell culture supernatants after knockdown of FX; p-ERK1/2 and p-AKT were increased when THP-1 cells were treated with U251 cell culture supernatants after overexpression of FX. (h) p-ERK1/2 and p-AKT increased in a concentration-dependent manner with the stimulation of recombinant FX (rFX). (i) THP-1 cells incubated with G1124-FX cell supernatants supplemented with or without p-ERK1/2 (GSK2606414) or p-AKT (MK2206) inhibitors. M1 macrophage marker levels increased, while M2 macrophage marker levels decreased with these two inhibitors (mean ± SEM, *p < 0.05, **p < 0.01). (J) p-ERK1/2 and p-AKT were detected by Western blotting in U251 and G1104 cell overexpression of FX and G1124 cell knockdown of FX.
growth in vivo, but did not affect GBM cell proliferation and invasion in vitro.
M1 TAMs play a role in antitumor immune responses, while M2 subtype TAMs are now widely regarded as immunosuppressive cells with a tumor-supportive role (38). CD11c is highly expressed in M1 subtype macrophages, and CD11c HI cells were more capable of protein antigen processing and associated with APC function (39). By contrast, CD163 is highly expressed in M2 subtype macrophages. CD163 binds hemoglobin-haptoglobin (Hb-Hp) complexes, which leads to endocytosis and degradation of Hb-Hp by heme-oxygenase enzymes, resulting in an anti-inflammatory response (40). Our studies confirmed that FX was secreted to the tumor microenvironment and influenced macrophage polarization to the M2 subtype. We also found that FX secreted to GBM cell supernatants decreased THP-1 cell M1 subtype marker (IL-1β, IL-12α, CXCL9, IL-12β, and CXCL10) expression, while increasing M2 subtype markers (LYVE1, MRC1, STAB1, and SerpinB2) expression. Moreover, FX increased the proportion of M2 subtype (CD11b + CD206 + ) but decreased M1 subtype (CD11b + CD80 + ) macrophages. IL-1β is the most studied pro-inflammatory cytokine in the IL-1 family and is crucial for inflammation and tissue damage (41). M1 macrophages highly express chemokines such as CXCL9 and CXCL10, which attract Th1 cells (42). IL-12 comprises p35 (encoded by IL-12α) and p40 (encoded by IL-12β) chains and is considered as a proinflammatory molecule, which principally activates natural killer cells and induces naïve CD4 + T lymphocytes to differentiate into Th1 effector cells (43). These molecules decreasing with the stimulation of FX suggested that FX may inhibit M1 subtype macrophage polarization to protect tumor cells from immune cells. By contrast, HMOX-1 has been recognized as having immunomodulatory and anti-inflammatory properties, which drive the phenotypic shift to M2 macrophages (44). ARG1 depleted l-arginine by metabolizing it to urea and l-ornithine. l-arginine is necessary for T cell function, and its depletion suppresses T cell activity (45). MRC1 is an innate pattern-recognition receptor and has functions in clearing allergens and limiting allergic inflammation (46). SerpinB2 represents an immune-regulated factor that is critical for macrophage survival (47). The increase in molecules increased with FX stimulation indicated that FX may promote macrophages toward the M2 subtype to establish an immunosuppressive environment that is beneficial for tumor growth.
Previous studies have suggested that FX bound to the αMβ2 integrin subunit CD11b on activated monocytes, which was responsible for the conversion of prothrombin to thrombin (48). Integrin was also involved in mediating cell adhesion and migration. Blocking integrin α1β1 or knocking out integrin subunit alpha 1 increased macrophage exit from inflamed skin, suggesting that integrin is critical for controlling macrophage recruitment (49). Our results further confirmed that the GBM cell-secreted FX that mediates TAM recruitment and polarization may via CD11b. Blocking of integrin αM subunits with the CD11b antibody prevented macrophage recruitment and M2 subtype macrophage polarization induced by FX.
Integrin has been demonstrated to mediate cell proliferation, differentiation, and migration via the ERK pathway (50). The clustering of integrin αMβ2 inhibits the apoptosis of human neutrophils by activating ERK and AKT (51). Integrin α2β1 promotes T cell migration via activation of the ERK/Mcl-1 and p38 MAPK pathways (52). In macrophages, phosphorylation and activation of MAPK have been implicated in regulating macrophage polarization (53). The activation and proliferation of macrophages require ERK1/2 phosphorylation, and sustained activation of MAPK phosphorylation increased the expression of TNF-α, IL-6, and IL-10 (54). Hypoxia promotes M2 subtype polarization via the activation of ERK and enhances metastasis in non-small cell lung cancer (55). ROS generation induces ERK activation, resulting in macrophage polarization to the M2 subtype (56). Fortunately, in our study, we first confirmed FX bound to ERK1/2 in the cytoplasm, and more importantly, GBM cells secreting FX promoted the phosphorylation and activation of ERK1/2, resulting in macrophage M2 subtype polarization. In addition, activation of the PI3K/AKT pathway is critical for suppressing pro-inflammatory responses, while promoting anti-inflammatory responses in macrophages (57). The PI3K-AKT and MEK1/2-ERK1/2 pathways often collaborate and cross talk with each other (58). The ERK inhibiter U0126 and β-adrenoceptor decrease ERK phosphorylation while increases AKT phosphorylation (59). Other studies showed that the PI3K inhibitor wortmannin reduced p-ERK, and the ERK inhibitor FR180204 also enhanced p-AKT (60). AKT activation is required for M2 activation by the upregulation of M2 genes and several molecules, such as TGF-β and IL-10 (61). TIPE2 promotes M2 macrophage polarization via activation of the AKT signaling pathway (62). In our study, FX secreted by GBM cells and increased p-AKT in macrophages to promote M2 macrophage polarization. But FX did not interact with AKT in GBM cells. Interestingly, inhibition of p-ERK1/2 and p-AKT eliminated the effects of FX on M2 macrophage polarization. Combined with the results of the CD11b antibody on macrophage polarization, we proposed that FX secreted from GBM cells to the tumor environment recruited macrophages by interacting with CD11b on the surfaces of macrophages. Furthermore, FX bound to CD11b to promote phosphorylation and activation of ERK1/2 and AKT, resulting in M2 macrophage polarization. It is well known that the MAPK/ERK and PI3K/AKT pathways are reportedly associated with cell proliferation, differentiation, migration, senescence, and apoptosis (63). In our study, we found that FX bound to ERK1/2 and decreased p-ERK1/2 in GBM cells. These results contradicted with that FX did not influence the proliferation and invasion of GMB cells. Other molecule may involved in FX-mediated cell proliferation and further studies needed to prove.
miR-338-3p is reported to function as a tumor suppressor gene in various cancer, including hepatocellular carcinoma, neuroblastoma, ovarian cancer, gastric cancer, and colorectal cancer (64)(65)(66). In this study, we found that miR-338-3p targeted FX by binding to its 3′UTR in GBM cells. Our recent study showed that the lncRNA CASC2c directly bound miR-101 and influenced miR-101 to mature, and CASC2c acted as a competing endogenous RNA (ceRNA) of miR-101 to competitively regulate CPEB1 and promote the malignant growth of astrocytoma (67,68). In this study, we first confirmed that in GBM cells, CASC2c repressed both miR-338-3p and FX. miR-338-3p bound CASC2c and inhibited CASC2c expression; on the other hand, CASC2c was a target gene of miR-338-3p and was repressed by miR-338-3p. Therefore, CASC2c and miR-338-3p formed a mutually inhibitory complex. However, CASC2c did not function as a ceRNA of miR-388-3p to competitively regulate FX expression. We found that CASC2c, along with miR-388-3p, inhibited FX expression. Yuan reported that the lncRNA ATB bound to IL-11 mRNA and increased IL-11 mRNA stability and secretion in hepatocellular carcinoma (69). Therefore, we proposed that CASC2c may bind FX mRNA directly and result in FX mRNA degradation. The mechanism needs to be studied further. In addition, both miR-338-3p and CASC2c in GBM cells repressed the migration of macrophages by inhibiting FX expression. CASC2c could influence macrophage polarization, but miR-338-3p could not. miR338-3p can target many proteins such as MMP2, MMP9 (70), BDNF (71), Sphk2 (72), and IRS2 (73). BDNF enriched in glioma environment stimulated the production of IL-15 in CD11b + macrophages cells and altered macrophage plasticity (74). Sphk2 was demonstrated as a novel driver of the proinflammatory macrophage phenotype (75). IRS2 negatively regulates YM1 protein induction in macrophages and negatively regulated alternative macrophage activation (76). Therefore, although miR338-3p targets FX, miR338-3p did not influence macrophage polarization. In conclusion, CASC2c repressed FX expression synergistically with miR-338-3p and CASC2c regulated macrophage polarization.
In summary, this study demonstrated a novel role of FX in GBM growth (Figure 8). FX was highly expressed in GBM cells, secreted to the tumor microenvironment and functioned as a potent chemokine in the chemoattraction of TAMs. FX recruited macrophages to promote tumor growth by interacting with CD11b on the surfaces of macrophages. FX promoted M2 subtype polarization by activating ERK1/2 and AKT in macrophages. In GBM cells, the lncRNA CASC2c interacted with and reciprocally repressed miR-338-3p; both CASC2c and miR-388-3p bound to FX, commonly inhibited the expression of FX. Our finding of TAM recruitment by FX may offer therapeutic potential to improve GBM therapy in patients.

eThics sTaTeMenT
This study was carried out in accordance with the recommendations of the Joint Ethics Committee of the Central South University Health Authority with written informed consent from all subjects. All animal experiments were performed in accordance with the guidelines for the care of laboratory animals and the Animal Care and Use Committee of Central South University.