Brucea Javanica Oil Attenuates Osteosarcoma Progression by Reprogramming Tumor-associated Macrophages via Inhibiting PI3K/AKT Pathway

Abstract

Osteosarcoma (OS) is a leading malignant bone tumor in children and adolescents. Resetting M2-polarized tumor-associated macrophages (TAMs) has emerged as a promising therapeutic target for OS. Brucea javanica has long been used in China for cancer treatment, but its effects on TAMs in OS remain underexplored. Human THP-1 and murine RAW264.7 macrophages were polarized to an M2 phenotype by IL-4 and/or IL-13. The effects of Brucea javanica oil (BJO) and acetone extract (DBA) on TAM repolarization were assessed by qRT-PCR, ELISA, and flow cytometry. BJO’s impact on M2-promoted OS cell proliferation, migration, invasion, and vasculogenic mimicry was evaluated in vitro. BJO were characterized by GC-TOF-MS analysis. Network pharmacology, western blot, and rescue experiments were carried out to uncover the mechanism through which BJO reprogram TAMs. An OS mouse model generated by co-implantation of K7M2 and M2-polarized RAW264.7 macrophages was used to evaluate the effects of BJO on OS growth and TAM reprogramming in vivo. BJO exhibits a stronger capacity than DBA to switch TAMs from M2 to M1 phenotype in a dose-dependent manner, thereby inhibiting M2 macrophage-promoted aggressive behaviors of OS cells in vitro. Nine constituents were identified in BJO by GC-TOF-MS. Mechanistically, BJO inhibits PI3K/AKT pathway in TAMs, thereby relieving its negative regulation of ERK1/2 and NF-κB signaling. In vivo, BJO inhibits OS growth and drives TAMs into an M1 phenotype, while demonstrating minimal systemic toxicity. These findings reveal a novel anti-OS mechanism of BJO, highlighting its potential as an M2-TAM-targeted immunotherapeutic agent.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s12010-026-05713-x.

Introduction

Osteosarcoma (OS), the most common primary malignant bone tumor, primarily affects children and adolescents, with a peak incidence between 10 and 20 years of age [1]. Despite its low annual incidence of approximately 2–5 cases per million individuals, OS imposes a substantial societal burden on pediatric populations [2]. Although therapeutic advancements have improved the 5-year survival rate from 20% to 65%, the subsequent plateau underscores the pressing need for innovative therapies [3]. Immunotherapy has emerged as a promising approach in cancer treatment by harnessing the body’s own immune system to target and destroy cancer cells [4]. Its efficacy offers hope for previously incurable cancers and opens new avenues for traditional cancer therapies [5]. In OS, the immune responses are often suppressed, thereby facilitating tumor progression and metastasis [6]. Immunotherapy seeks to overcome immunosuppression by bolstering the body’s ability to recognize and eliminate OS cells [7]. Although immunotherapy has not been routinely adopted in the clinical management of OS, its established efficacy in other malignancies suggests it represents a potential therapeutic avenue for OS, warranting further investigation [7].

The OS tumor microenvironment (OS-TME) critically influences tumor progression and therapeutic resistance through its orchestration of interactions between malignant cells and stromal components [6]. This ecosystem includes various immune cell types [5], among which tumor-associated macrophages (TAMs) are the most abundant, constituting approximately 50% of tumor-infiltrating leukocytes [8]. The remarkable phenotypic plasticity of TAMs plays a critical role in the pathogenesis of OS [9]. Within OS lesions, TAMs exist in a dynamic balance between M1 and M2 polarization states. Th2 cytokines (IL-4/IL-13) drive M2 polarization, characterized by the secretion of immunosuppressive factors such as IL-10 and TGF-β [10]. Conversely, TLR agonists and IFN-γ induce M1 polarization [11]. M1-TAMs exert anti-tumor activity via tumor cell phagocytosis, antigen presentation, and the release of cytotoxic molecules, whereas M2-TAMs facilitate tumor progression through promoting angiogenesis and enhancing immune checkpoint-mediated immunosuppression [12]. Clinical observations indicate that TAMs within the OS-TME are predominantly skewed toward an M2 phenotype [13]. A higher M2/M1 ratio is associated with increased metastasis and poorer overall survival [13]. In light of the pro-tumorigenic functions of M2-TAM, redirecting their polarization toward the M1 phenotype through pharmacological intervention has gained attention as a therapeutic approach for OS [7].

Brucea javanica is a traditional Chinese herbal medicine first documented in the sixteenth-century Chinese medical monograph The Omissions from the Compendium of Materia Medica. The drug source of Brucea javanica is derived from the dried ripe fruit of Brucea javanica (L.) Merr [14]., which has historically been used to treat cancer due to its heat-clearing and detoxifying properties [15]. These properties are attributed to its cold and bitter nature [15]. Brucea javanica ethanol extracts contain brusatol and bruceine D, two bioactive constituents with potent anti-tumor activity [16–17]. Moreover, Brucea javanica oil (BJO), a clinically applied extract, has been reported to increase clinical response rates in patients with lung and gastrointestinal cancer [18]. Its ability to modulate the tumor immune microenvironment has attracted growing attention, as preclinical evidence suggest it can augment the efficacy of anti-PD-1 immunotherapy [19]. Nonetheless, whether TAMs in OS can be effectively repolarized by Brucea javanica extract, and which specific fractions are responsible for this process remains unclear. Therefore, this study investigated the ability of Brucea javanica fractions to repolarize M2-TAMs towards an M1 phenotype and evaluated their consequent anti-OS effects.

Materials and Methods

Acquisition of DBA and BJO

Brucea javanica herbal material was sourced from the Jiangsu Province Academy of Traditional Chinese Medicine and authenticated by Professor Li Zhenlin as the dried mature fruits of Brucea javanica (L.) Merr. (family Simaroubaceae). As previously reported [20], a 500 g portion of the Brucea javanica air-dried fruits was extracted using the Soxhlet extraction with 60–90% petroleum ether as the solvent. The extraction was performed twice at 85 °C for 1.5 h each time with a liquid-to-material ratio of 8:1 (mL: g), and the combined extracts were concentrated under reduced pressure to yeild 63.2 g of BJO. The residual material was further extracted with acetone under reduced pressure at 75 °C, using a liquid-to-solid ratio of 6:1 (mL: g) for 1.5 h. This extraction process was repeated twice, affording 49 g of defatted Brucea javanica acetone extract (DBA).

Reagents

Interleukin-4 (IL-4) and Interleukin-13 (IL-13) were procured from PeproTech (Rocky Hill, NJ, USA). Phorbol 12-myristate 13-acetate (PMA) was purchased from Sigma-Aldrich (St. Louis, MO, USA). RPMI-1640, DMEM, trypsin, phosphate-buffered saline (PBS), and penicillin-streptomycin (PS) were purchased from KeyGen Biotech (Nanjing, China). Fetal bovine serum (FBS) was obtained from Gibco (Carlsbad, CA, USA). MTT, RIPA, crystal violet, bicinchoninic acid (BCA) assay, protein quantification kit, paraformaldehyde (PFA), and nitric oxide (NO) assay kits were purchased from Beyotime Biotech (Beijing, China). TRIzol reagent was purchased from Vazyme (Nanjing, China). The ECL kit was purchased from Wanleibio (Shenyang, China). Phenol red-free Matrigel matrix and Transwell inserts were purchased from Corning (NY, USA). Alanine aminotransferase (ALT), aspartate aminotransferase (AST), urea (BUN), and creatinine (CR) assay kits were obtained from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Enzyme-linked immunosorbent assay (ELISA) kits for human TNF-α, human IL-10, mouse TNF-α, and mouse IL-10 were obtained from MultiSciences Biotech (Hangzhou, China), while kits for human ARG1 and mouse ARG1 were obtained from Elabscience (Wuhan, China). Primary antibodies against p-p38 (#28796-1-AP), p38 (#14064-1-AP), p-ERK1/2 (#28733-1-AP), ERK1/2 (#11257-1-AP), p-JNK (#80024-1-RR), JNK (#66210-1-Ig), p-IκB (#82349-1-RR), IκB (#10268-1-AP), and GAPDH (#10494-1-AP), as well as secondary antibodies including HRP-conjugated goat anti-rabbit (#SA00001-2) and goat anti-mouse (#SA00001-1) IgG, were obtained from ProteinTech (Wuhan, China). The anti-CD86 (#abs120515) antibody was purchased from Absin Bioscience (Shanghai, China). Antibodies against p-NF-κB p65 (#8242), NF-κB p65 (#3033), PI3K (#4292), AKT (#9272), p-AKT (#2974T), mTOR (#2972), p-mTOR (#9271T), CD86 (#19589T) and CD206 (#24595T) were purchased from Cell Signaling Technology (Danvers, MA, USA). The anti-F4/80 (#ab300421) antibody was purchased from Abcam (Cambridge, UK). The anti-p-PI3K (#YP0224) antibody was purchased from ImmunoWay Biotechnology (Plano, TX, USA). The anti-β-tubulin (#WL01931) antibody was purchased from Wanleibio (Shenyang, China). FITC-conjugated anti-human CD86 (#374203), anti-human CD206 (#321103), anti-mouse CD86 (#159219), and anti-mouse CD206 (#141703) antibodies were purchased from BioLegend (San Diego, CA, USA). Phosphatase inhibitor cocktail, U0126, BAY 11-7082, and 740 Y-P were purchased from MedChemExpress (Shanghai, China). Fluorescein tyramide, CY3 tyramide and CY5 tyramide were purchased from AAT Bioquest (Pleasanton, CA, USA). MaxVision™ HRP-Polymer anti-Rabbit IHC Kit were purchased from MXB Biotechnologies (Fuzhou, China).

Sample Preparation

An aliquot of 3 mL of BJO was saponified with 2 mL of 0.5 M methanolic potassium hydroxide at 60 °C for 25 min. After cooling, methylation was performed by adding 2 mL of a 15% boron trifluoride-diethyl ether complex at 60 °C for 2 min. The extract was then extracted by adding 2 mL of n-hexane and 1 mL of saturated NaCl solution. The hexane layer was collected for GC-TOF-MS analysis.

GC-TOF-MS Analysis

Mass spectrometry data were acquired using an Agilent 7250 GC-QTOF high-resolution mass spectrometer equipped with a HP-5MS Ultral I (15 m × 0.25 mm × 0.25 μm, Agilent J&W, Folsom, CA, USA). Sample was injected splitlessly at 280 °C with purge flow rate of 3 mL/min, using helium as carrier gas at a flow rate of 1.0 mL/min. The injection volume was 1 µL. The GC oven temperature program began at 40 °C (held for 1 min), then increased to 280 °C at a rate of 20 °C/min. MS employed an EI ionization source (70 eV) with an ion source temperature of 220 °C. Mass-to-charge ratio (m/z) was scanned from 50 to 500 at an acquisition rate of 20 spectra/s, with the transfer line maintained at 270 °C.

Cell Culture

The human monocytic leukemia cell line THP-1 (Official name: THP-1. RRID: CVCL_0006.), the murine macrophage cell line RAW264.7 (Official name: RAW 264.7. RRID: CVCL_0493), and the human OS cell line 143B (Official name: 143B. RRID: CVCL_2270) were obtained from KeyGen Biotech. The mouse OS cell line K7M2 (Official name: K7M2 wt. RRID: CVCL_V455) was purchased from the Shanghai Institute of Cell Biology (Shanghai, China). THP-1 cells were maintained in RPMI-1640 supplemented with 10% FBS and 1% PS. RAW264.7 and K7M2 cells were cultured in DMEM with 10% FBS and 1% PS, whereas 143B cells were cultured in MEM with 10% FBS and 1% PS. Cells were maintained at 37 °C with 5% CO2 in a humidified incubator.

Macrophage Polarization and CM Acquisition

THP-1 cells were stimulated with 50 ng/mL PMA for 48 h to induce differentiation into M0 macrophages, followed by polarization into M2 macrophages using 10 ng/mL IL-4 and 20 ng/mL IL-13 for 48 h. RAW264.7 cells were polarized toward the M2 phenotype by stimulation with 50 ng/mL IL-4 for 48 h. After polarization, cells were washed and cultured in fresh medium to generate M2 macrophage-conditioned medium (M2-CM). BJO and DBA were dissolved in DMSO (final concentration < 0.1% v/v) and subsequently diluted with culture medium. M2 macrophages were treated with BJO -containing medium for 48 h. The supernatant was then harvested, and designated as M2-BJO-CM. The supernatants from THP-1-derived M0 macrophages and RAW264.7 cells were collected and designated as M0-CM.

MTT assay

Macrophages or OS cells were seeded in 96-well plates and treated with either BJO or DBA at varying concentrations or with conditioned media, including M2-BJO-CM (at different BJO concentrations, with or without phosphatase inhibitors/agonists) or M0-CM. After 48 h of treatment, MTT solution was added and incubated for an additional 4 h. Formazan crystals were dissolved in DMSO, and absorbance was recorded at 490 nm using a microplate reader (Synergy H1, BioTek, USA).

Quantitative PCR (qPCR)

Cells or OS mouse tumor samples were lysed in Trizol, and complementary cDNA was synthesized using the All-in-One First-Strand Synthesis MasterMix Version 2 kit (#EG15133S, Yugong Biolabs, Lianyunagang, China) according to the manufacturer’s instructions. qPCR was performed using Taq SYBR^® Green qPCR Premix (#EG20117M, Yugong Biolabs) on an Applied Biosystems QuantStudio 7 Real-Time PCR System (Thermo Fisher, USA). GAPDH was used as an internal control for the gene expression analysis. The sequences of primers used are listed in Table 1.

Table 1.

Primer sequences of qPCR

Gene	Forward Primers 5′-3′	Reverse Primers 5′-3′
Mouse
Gapdh	GTGGCAAAGTGGAGATTGTTG	CGTTGAATTTGCCGTGAGTG
Tnf	CAGGCGGTGCCTATGTCTC	CGTCACCCCGAAGTTCAGTAG
Il1b	GAAATGCCACCTTTTGACAGTG	TGGATGCTCTCATCAGGACAG
Ccl22	CAGGTCCCTATGGTGCTAATG	GGAGGGTTTCTGGGTTTTCTT
Il6	TAGTCCTTCCTACCCCAATTTCC	TTGGTCCTTAGCCACTCCTTC
Nos2	TGCCACGGACGAGACGGATAG	CTCTTCAAGCACCTCCAGGAACG
Il10	CTTACTGACTGGCATGAGGATCA	GCAGCTCTAGGAGCATGTGG
Arg1	CAAGACAGGGCTCCTTTCAG	TGGCTTATGGTTACCCTCCC
Mrc1	CTCTGTTCAGCTATTGGACGC	TGGCACTCCCAAACATAATTTGA
Human
GAPDH	TGGCTTATGGTTACCCTCCC	CTGGAAGATGGTGATGGGATT
TNF	CCTCTCTCTAATCAGCCCTCTG	GAGGACCTGGGAGTAGATGAG
IL1B	AATCTCACAGCAGCATCTCGACAAG	TCCACGGGCAAGACATAGGTAGC
CCL22	TTCCTGCTGTTTCTGCCTCAC	TGGGTGATGTTGAACTCAGCA
IL6	ACTCACCTCTTCAGAACGAATTG	CCATCTTTGGAAGGTTCAGGTTG
NOS2	ACATCGACCCGTCCACAGTAT	CAGAGGGGTAGGCTTGTCTC
IL10	GACTTTAAGGGTTACCTGGGTTG	TCACATGCGCCTTGATGTCTG
ARG-1	GTGGAAACTTGCATGGACAAC	AATCCTGGCACATCGGGAATC
MRC1	GGGTTGCTATCACTCTCTATGC	TTTCTTGTCTGTTGCCGTAGTT

Flow Cytometry

M0, untreated M2, and BJO or DBA-treated M2 macrophages were harvested by trypsinization, collected, and centrifuged at 1000 rpm for 5 min. The cell pellet was washed twice with ice-cold PBS. Cells were resuspended in 90 µL of PBS and incubated with FITC-labelled anti-human CD86, anti-human CD206, anti-mouse CD86, or anti-mouse CD206 antibodies. After incubation at 37 °C for 30 min in the dark, cells were washed with 400 µL of PBS, centrifuged, and resuspended in 300 µL of PBS for immediate analysis using a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA). The proportion of positive cells was analyzed using FlowJo software (Tree Star, Ashland, OR, USA).

ELISA and NO Assay

ARG1, TNF-α, IL-10 and NO levels in macrophage supernatants or homogenized OS tissue samples from mice were determined using commercial ELISA kits or NO assay kits, following the manufacturer’s protocols.

Colony Formation Assay

K7M2 or 143B cells were seeded in six-well plates at a density of 5 × 10² cells per well and incubated overnight. Cells were then treated with various CM for one week, fixed with 4% PFA, stained with crystal violet, visualized under a light microscope (Eclipse TiU, Nikon, Japan) and quantified using ImageJ software (NIH, USA).

Wound Healing Assay

K7M2 or 143B cells were cultured in six-well plates. After cell adherence, a scratch was made using a pipette tip, followed by incubation with various CM. Images were acquired at 48 h using a light microscope and analyzed using ImageJ.

Migration and Invasion Assay

K7M2 or 143B cells were seeded in the upper chambers of 24-well Transwell inserts, with various CM added to the lower chambers. Inserts were either coated with Matrigel for invasion assays or left uncoated for migration assays. After 48 h of incubation, cells on the upper surface were removed, and those on the lower surface of the membrane were fixed with 4% PFA, stained with crystal violet, visualized under a light microscope and quantified using ImageJ.

Tube Formation Assay

A 24-well plate was coated with 200 µL of Matrigel. After gel solidification, K7M2 or 143B cells were resuspended in various CM and seeded into each well. After an 8 h incubation, tube formation was visualized under a light microscope. Quantitative analysis of the vascular network was performed with the ‘Angiogenesis Analyzer’ plugin of ImageJ.

Network Pharmacology Analysis

The primary bioactive compounds in BJO were determined by combining the compounds identified by GC-TOF-MS analysis with those predicted via TCMSP screening. For the TCMSP screening, “Brucea javanica” was used as the search keyword with the following criteria: oral bioavailability (OB) ≥ 35% and drug-likeness (DL) ≥ 0.18. The potential targets of the combined compounds were systematically identified using the Swiss Target Prediction Database. Targets associated with OS were retrieved from the GeneCards and PharmGKB databases, whereas those related to TAM were obtained from the GeneCards and OMIM databases. Venn diagram analysis was conducted using the VENNY 2.1 platform to determine the overlapping targets among the datasets. The shared targets were subsequently uploaded to the STRING database, where the species type was set to “Homo sapiens” and the minimum interaction threshold was set to “medium confidence” (> 0.4). The protein-protein interaction (PPI) network and the Pathway-Target-Compound network were imported into Cytoscape 3.9.1 software for visual analysis. Topological analysis of the PPI network was performed using the CytoNCA plugin in Cytoscape. The top 15 hub genes were identified based on degree centrality, and node colors were subsequently adjusted for visualization. Finally, KEGG pathway enrichment analysis was performed using the Sangerbox online platform with the species set to Homo sapiens, a P-value cutoff of 0.01, and a minimum gene count of three.

Western Blot

Following homogenization in ice-cold RIPA buffer containing protease and phosphatase inhibitor cocktails, cell lysates were centrifuged, and protein levels was quantified using the BCA method. Following SDS-PAGE, proteins were transferred onto nitrocellulose membranes (Millipore, Billerica, MA, USA). The membranes were immunostained with specific primary antibodies overnight at 4 °C, and subsequently incubated with secondary antibodies. Chemiluminescent signals were captured by ECL system, and band intensities were semi-quantitatively analyzed with ImageJ.

K7M2 Xenografts in Mice

Animal use was approved by the Ethics Committee of Jiangsu Province Academy of Traditional Chinese Medicine (Approval No. AEWC-20240202-368) and adhered to the established protocols for the ethical use and care of laboratory animals. BALB/C mice (6 weeks old) were purchased from Suzhou SiPeiFu Bio-Tech Co, Ltd (SCXK (HU) 2022-0006). Animals were maintained under pathogen-free conditions, with temperature (22 ± 1 °C), humidity (50 ± 1%), and a 12 h light/dark cycle strictly regulated. They were provided unrestricted access to food and water and underwent a 7-day acclimation period prior to experiment.

Mice were randomly divided into four groups (8 mice/group): K7M2 group, K7M2 + M2 group, K7M2 + BJO group and K7M2 + M2+BJO group. For the syngeneic tumor model, cells suspended in 0.1 mL of PBS were subcutaneously injected into the right axillary region of the mice. Mice in the K7M2 group and K7M2 + BJO group were injected with 4 × 10⁶ K7M2 cells. The mice in the K7M2 + M2 group and K7M2 + M2+BJO group received a mixture of 4 × 10⁶ K7M2 cells and 6 × 10⁶ IL-4/13-treated RAW264.7 cells. Tumor dimensions were recorded every two days, and volume was estimated as Volume = (Length × Width²)/2. Once the tumor size reached ~ 100 mm³, mice received daily oral gavage of BJO (400 mg/kg) dissolved in corn oil or corn oil alone (10 mL/kg) as a control. An additional blank control group was included, receiving purified water by daily oral gavage.

After three weeks of treatment, mice were humanely sacrificed by inhalation anesthesia with isoflurane. Blood specimens were acquired through retro-orbital puncture for serum and plasma isolation. Mouse tumor tissues and major organs were harvested. Tumors were partitioned, with one portion fixed in 4% PFA for further haematoxylin-eosin (H&E) staining, immunohistochemistry, and immunofluorescence, while the remaining portions were stored at -80 °C for qPCR and ELISA.

Histopathological Analysis

Tumor tissues fixed in 4% PFA were paraffin-embedded, sectioned, and subjected to H&E for histopathological evaluation.

Immunohistochemical Staining

For immunohistochemical staining, the sections were incubated with anti-Ki-67 antibody, visualized using the labeled streptavidin-biotin detection system, and counterstained with hematoxylin. Semi-quantitative assessment was carried out using ImageJ.

For CD31/PAS double staining, sections were initially incubated with anti-CD31 antibody, followed by treatment with 0.5% periodic acid solution. Quantitative analysis of the double-stained sections was performed as described previously [21]. Images were acquired using a light microscope.

Immunofluorescence Staining

Tissue sections were deparaffinized and rehydrated, followed by microwave-assisted antigen retrieval in citrate buffer. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide, and nonspecific binding sites were blocked with 3% BSA. For Phase 1 staining, sections were incubated overnight at 4 °C with primary rabbit monoclonal antibodies against CD86 or CD206. After washing, bound antibodies were detected using a MaxVision™ HRP-polymer anti-rabbit IHC kit, followed by tyramide signal amplification with either CY3 tyramide (for CD86) or CY5 tyramide (for CD206). Phase 2 staining was performed by incubating sections with a rabbit anti-F4/80 primary antibody overnight at 4 °C. The same MaxVision™ HRP-polymer system was applied, followed by signal amplification using fluorescein tyramide. Nuclei were counterstained with DAPI. Fluorescence images were then acquired using a fluorescence microscope (Olympus VS200, Tokyo, Japan). The positive signals were quantified using ImageJ.

Statistical Analysis

Data are expressed as mean ± standard deviation (SD) from at least three independent replicates for in vitro experiments and at least five independent replicates for in vivo experiments. Statistical comparisons between two groups were performed using Student’s t-test, while differences among multiple groups were analyzed by one-way analysis of variance (ANOVA) followed by Dunnett’s multiple comparisons test.

All analyses were conducted using GraphPad Prism 9.5 software (San Diego, CA, USA). Differences were considered statistically significant when P value < 0.05.

Results

BJO Drives M2-to-M1 Macrophage Repolarization In Vitro

In our preliminary experiment, we established a M2 macrophage polarization model from THP-1 cells using a classic induction protocol with IL-4 and IL-13 [11, 22]. Successful polarization was confirmed by a marked increase in the transcription of M2 markers (e.g., MRC1 and IL10) and a concurrent decrease in M1 markers (e.g., TNF and IL1B) in M2 macrophages relative to M0 cells [10]. Consistently, the proportion of CD206⁺ cells increased from 5.79% in the M0 group to 65.88% in the M2 group (Fig. S1A). Similarly, RAW264.7 murine macrophages were successfully polarized toward an M2 phenotype by a 48-hour IL-4 stimulation protocol [23]. At the transcriptional level, M2 macrophages exhibited significantly elevated expression of M2 markers (Mrc1 and Il10) and reduced expression of M1 markers (Tnf and Il1b) compared with M0 macrophages. In parallel, flow cytometry revealed a substantial increase in the CD206⁺ macrophage population from 6.19% in M0 cells to 68.35% in M2 cells (Fig. S1B).

Before examining the effects of BJO and DBA on M2 phenotypic transition, we first evaluated their cytotoxicity on M2-polarized THP-1 and RAW264.7 cells using the MTT assay. As shown in Fig. 1A-B, neither fraction treatment significantly inhibited the growth of these M2-polarized macrophages at concentrations below 1 µL/mL. Accordingly, this concentration range was applied in the subsequent experiments. Next, we analyzed the expression of M1 and M2 marker genes by qPCR. As shown in Fig. 1C-D, both fractions suppressed M2 macrophage polarization, but with distinct potencies. BJO induced a more pronounced phenotypic shift, significantly upregulating M1 markers—including TNF, IL6, NOS2, and IL-1B-and robustly downregulating M2 markers-including IL10, MRC1, ARG1, and CCL22—in M2-polarized THP-1 and RAW264.7 macrophages. In contrast, DBA only significantly increased TNF mRNA mRNA levels and elicited a more modest downregulation of M2-associated genes.

Fig. 1.

BJO induces the repolarization of M2 macrophages toward the M1 phenotype in vitro. (A-B) Cytotoxic effects of BJO or DBA on M2 macrophages polarized from THP-1 and RAW264.7 cells following 48 h exposure, as assessed by MTT assay (n = 6). (C-D) qPCR analysis of mRNA expression levels of TNF, IL6, NOS2, IL1B, IL10, MRC1, ARG1, and CCL22 in M2 macrophages polarized from THP-1 cells, and the corresponding murine genes (Tnf, Il6, Nos2, Il1b, Il10, Mrc1, Arg1, and Ccl22) in M2 macrophages polarized from RAW264.7 cells following 48 h exposure to BJO (1 µL/mL) or DBA (1 µL/mL) (n = 3). (E-F) Flow cytometry analysis of CD86 and CD206 expression in M2 macrophages polarized from THP-1 and RAW264.7 cells following 48 h exposure to BJO (1 µL/mL) or DBA (1 µL/mL) (n = 3). (G) qPCR analysis of mRNA expression levels of TNF, IL6, NOS2, IL1B, IL10, MRC1, ARG1, and CCL22 in M2 macrophages polarized from THP-1 cells and the corresponding murine genes (Tnf, Il6, Nos2, Il1b, Il10, Mrc1, Arg1, and Ccl22) in M2 macrophages polarized from RAW264.7 cells following 48 h exposure to increasing concentrations of BJO (n = 3). (H) Concentrations of TNF-α, NO, IL-10, and ARG1 in conditioned media from M2 macrophages polarized from THP-1 and RAW264.7 cells after 48 h exposure to increasing concentrations of BJO, quantified by ELISA or an NO assay kit (n = 3). Data are expressed as the mean ± SD of at least three independent experiments. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 vs. M2 or vehicle control group

Consistent with the qPCR results, flow cytometry analysis showed BJO treatment increased the proportion of CD86⁺ cells from 6.99% to 39.66% and reduced that of CD206⁺ cells from 65.88% to 26.96% in IL-4/IL-13-induced THP-1 macrophages. Similar effects were observed in IL-4-induced RAW264.7 macrophages, where CD86⁺ cells rose from 10.6% to 40.9% while CD206⁺ cells declined from 68.35% to 27.96%. Conversely, DBA induced only modest shifts in CD86⁺ and CD206⁺ populations, with markedly weaker effects than those observed with BJO in both cell lines (Fig. 1E–F). Collectively, these results indicate that BJO exhibits a markedly stronger capacity than DBA to repolarize M2 macrophages toward an M1 phenotype.

The dose-dependent effects of BJO (0.25-1 µL/mL) on the repolarization of M2 macrophages were further investigated. qPCR analysis revealed a concentration-dependent shift in polarization, with progressively increased mRNA levels of the aforementioned M1-associated markers and decreased expression of M2-associated markers (Fig. 1G). Consistently, ELISA and biochemical assays revealed that BJO treatment dose-dependently enhanced the secretion of TNF-α and production of NO, while concomitantly reducing IL-10 secretion and Arg-1 activity (Fig. 1H). Together, these findings indicate that BJO drives the phenotypic reprogramming of M2-polarized macrophages toward an M1-like state in a concentration-dependent manner.

BJO Reprograms M2 Macrophages to Suppress OS Cell Aggressiveness In Vitro

Accumulating evidence indicates that M2 macrophages play a key role in promoting malignant phenotypes in cancer cells [9]. We therefore investigated whether BJO could attenuate the pro-tumorigenic effects mediated by M2 macrophages. Direct treatment with BJO at concentrations below 2 µL/mL did not affect the viability of 143B and K7M2 cells (Fig. 2A). CM from M0, M2, and BJO-treated M2 macrophages—derived from THP-1 and RAW 264.7 cells—was applied to OS cells. CM derived from M2 macrophages significantly enhanced the proliferation of 143B and K7M2 cells by 71.17% and 74.67%, respectively, compared to M0-CM. This proliferative effect was dose-dependently suppressed by M2-BJO-CM (0.25-1 µL/mL) (Fig. 2B). Consistently, colony formation assays demonstrated that M2-CM markedly enhanced the long-term clonogenic potential of OS cells, an effect that was notably attenuated by M2-BJO-CM (Fig. 2C). Similarly, M2-CM promoted OS cell migration and invasion, whereas M2-BJO-CM substantially reduced these pro-metastatic behaviors. M2 macrophages have been implicated in promoting vasculogenic mimicry (VM), a malignant trait associated with strong aggressiveness in cancer cells [24]. After 6 h on Matrigel, 143B and K7M2 cells treated with M2-CM exhibited 47.36% and 49.02% increases, respectively, in capillary-like structures compared with the M0-CM group. In contrast, treatment with M2-BJO-CM reduced branch point formation in a dose-dependent manner by 38.28%, 60.23%, and 81.26% in 143B cells, and 30.79%, 47.36%, and 74.17% in K7M2 cells at BJO concentrations of 0.25, 0.5, and 1 µL/mL, respectively. Notably, direct exposure of OS cells to BJO did not alter their proliferation, migration, invasion, and VM formation, indicating that the observed anti-OS effects of BJO are mediated through its reprogramming of M2 macrophages rather than through direct action on OS cells.

Fig. 2.

BJO-mediated repolarization of M2 macrophages attenuates the malignant phenotype of OS cells in vitro. (A) Cytotoxic effects of BJO at various concentrations on 143B and K7M2 cells (n = 6). ^***p < 0.001 vs. vehicle control group. (B-G) 143B or K7M2 cells were exposed to conditioned media from THP-1 or RAW264.7-derived M0, M2, or BJO-treated M2 macrophages, respectively, or treated with BJO alone. (B) Cell viability assay in 143B and K7M2 cells (n = 3). (C) Colony formation assay in 143B and K7M2 cells (n = 3). (D) Wound healing assay in 143B and K7M2 cells (n = 3). Scale bar = 100 μm. (E-F) Migration and invasion assay in 143B and K7M2 cells (n = 3). Scale bar = 100 μm. (G) Vasculogenic mimicry assay in 143B and K7M2 cells (n = 3). Scale bar = 100 μm. Data are expressed as the mean ± SD of at least three independent experiments. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 vs. M2-CM group; ^##p < 0.01, ^###p < 0.001 vs. M0-CM group

GC-TOF-MS Analysis of the Chemical Composition of BJO

GC-TOF-MS analysis was employed to characterize the chemical composition of BJO, aiming to identify the bioactive constituents responsible for its ability to reprogram M2 macrophage (Fig. 3). The acquired spectra were compared against the NIST 20.0 mass spectral library, and only compounds with a Reverse Similarity Index (RSI) greater than 80 were considered confidently identified [25]. These putative identifications were further cross-referenced with established chemical databases, including ChemSpider and PubChem. This process led to the identification of nine major compounds in BJO, whose structural formulas and relevant information are summarized in Table 2, thereby laying the groundwork for functional and mechanistic studies.

Fig. 3.

Base peak chromatogram of BJO (1, Diethyl phthalate; 2, 3-Methylbutyl benzoate; 3, 9-Methylene-9 H-fluorene; 4, 2-Methylbutyl 4-methylbenzoate; 5, Methyl palmitate; 6, Methyl linoleate; 7, Methyl oleate; 8, Methyl stearate and 9, Hentriacontane)

Table 2.

GC-TOF-MS detection data of BJO after methyl esterification

NO.	Compound name	CAS	Formula	Rt (min)	RSI
1	Diethyl phthalate	84-66-2	C12H14O4	9.5318	90.2805
2	3-Methylbutyl benzoate	41757-90-8	C12H16O2	10.1975	80.7724
3	9-Methylene-9 H-fluorene	4425-82-5	C14H10	10.6697	88.8916
4	2-Methylbutyl 4-methylbenzoate	99-94-5	C13H18O2	10.8552	83.8465
5 6	Methyl palmitate Methyl linoleate	112-39-0 112-63-0	C17H34O2 C19H34O2	11.2889 12.1246	89.6732 86.2523
7	Methyl oleate	112-62-9	C19H36O2	12.1473	93.3422
8	Methyl stearate	112-61-8	C19H38O2	12.2571	81.3302
9	Hentriacontane	630-04-6	C31H64	13.0138	82.0290

CAS CAS registry number; Rt retention time; RSI reverse similarity index

BJO Reprograms M2 Macrophages by Inhibiting the PI3K/AKT Pathway

To elucidate the molecular mechanisms by which BJO repolarizes M2 macrophages, we first employed an integrated network pharmacology approach. From the TCMSP repository, we retrieved 67 known bioactive constituents of BJO, which we filtered using thresholds of OB ≥ 30% and DL ≥ 0.18, yielding 17 compounds that met these criteria. In addition, we incorporated the nine compounds identified in BJO by GC-TOF-MS analysis. After merging and removing duplicates, a total of 26 active components were chosen for subsequent analysis. Putative targets of these BJO components were retrieved from the SwissTargetPrediction database, resulting in 570 targets. After deduplication, 4880 OS-related genes were sourced from the GeneCards and pharmGKB databases, and 3032 genes associated with TAM were obtained from the GeneCards and OMIM databases. Venn diagram analysis identified 211 common molecular targets (Fig. 4A). To investigate interactions among these 211 targets, a PPI network was constructed, encompassing 190 nodes and 7504 edges (Fig. 4B). Topological analysis identified AKT1, TNF, and EGFR as core hub genes, with each ranking among the top 15 for connectivity (Fig. 4C). To investigate the underlying mechanisms associated with the potential targets of BJO, we performed KEGG pathway enrichment analysis. Among the top 10 significantly enriched pathways, several were cancer-related, including pathways in cancer, proteoglycans in cancer, prostate cancer, EGFR tyrosine kinase inhibitor resistance, and central carbon metabolism in cancer. Notably, the PI3K-Akt signaling pathway ranked second in enrichment significance and is a key pathway implicated in M2 macrophage polarization [25] (Fig. 4D). Next, a compound-target-pathway network was built in Cytoscape, with green nodes representing BJO-derived active compounds, red nodes their corresponding targets, and blue nodes signaling pathways (Fig. 4E).

Fig. 4.

Network pharmacology analysis of BJO. (A) Venn diagram of overlapping targets among BJO active compounds, TAM-related genes, and OS-related genes. (B) PPI network of 209 overlapping targets displayed in a spiral layout and ranked by their degree values. (C) Top 15 hub genes in the PPI network identified by CytoNCA plugin based on degree values, with node color ranging from light yellow to red representing increasing connectivity. (D) KEGG enrichment analysis of the overlapping targets. (E) Visualization of the compound-target-pathway network. Green rectangles indicate BJO active ingredients, red rectangles indicate the overlapping targets, and blue rectangles indicate enriched pathways

To validate these in silico predictions, we performed western blot analysis to assess the activation status of the PI3K/AKT/mTOR pathway. As shown in Fig. 5A, the phosphorylation levels of PI3K, AKT, and mTOR were significantly elevated in M2 macrophages compared to the M0 macrophages, a trend that was effectively reversed by BJO treatment. Concurrently, BJO administration decreased CD206 and increased CD86 protein expression levels. Given the well-established roles of MAPK and NF-κB signaling pathways in macrophage repolarization and inflammatory regulation, we assessed their activation in response to BJO treatment. Among the three MAPK members (p38, ERK1/2, and JNK), only phosphorylation of ERK1/2—but not that of p38 or JNK—was enhanced in both M2-polarized THP-1 and RAW264.7 cells following BJO treatment (Fig. 5B and Fig. S2A-B), implicating ERK1/2 in BJO-driven macrophage phenotype switching. Then, we found that BJO treatment increased phosphorylation of NF-κB p65 and IκB-α, along with a reduction in total IκB-α protein levels, suggesting activation of NF-κB signaling pathway in M2-polarized THP-1 and RAW264.7 cells (Fig. 5C).

Fig. 5.

BJO reprograms M2 macrophages by inhibiting the PI3K/AKT pathway. (A) The protein expression levels of p-PI3K, PI3K, p-AKT, AKT, p-mTOR, mTOR, CD86, and CD206 in THP-1 or RAW264.7-derived M0, M2 and M2 macrophages treated with the indicated concentrations of BJO were analyzed by western blot assay (n = 3). (B) The protein expression levels of p-ERK1/2 and ERK1/2 in THP-1 or RAW264.7-derived M0, M2 and M2 macrophages treated with the indicated concentrations of BJO were analyzed by western blot assay (n = 3). (C) The protein expression levels of p-NF-κB p65, NF-κB p65, p-IκB, and IκB in THP-1 or RAW264.7-derived M0, M2 and M2 macrophages treated with the indicated concentrations of BJO were analyzed by western blot assay (n = 3). (D) The protein expression levels of p-AKT, AKT, p-mTOR, mTOR, CD86, and CD206 in THP-1 or RAW264.7-derived M0, M2 and M2 macrophages pretreated with or without 740 Y-P prior to BJO treatment (1 µL/mL) were analyzed by western blot assay (n = 3). (E) The protein expression levels of p-ERK1/2, ERK1/2, CD86 and CD206 in THP-1 or RAW264.7-derived M0, M2 and M2 macrophages pretreated with or without U0126 prior to BJO treatment (1 µL/mL) were examined by western blot assay (n = 3). (F) The protein expression levels of p-NF-κB p65, NF-κB p65, p-IκB, IκB, CD86 and CD206 in THP-1 or RAW264.7-derived M0, M2 and M2 macrophages pretreated with or without BAY 11-7082 prior to BJO treatment (1 µL/mL) were examined by western blot assay (n = 3). (G) The protein expression levels of p-ERK1/2 and ERK1/2 in THP-1 or RAW264.7-derived M0, M2 and M2 macrophages pretreated with or without 740 Y-P prior to BJO treatment (1 µL/mL) were examined by western blot assay (n = 3). (H) The protein expression levels of p-NF-κB p65, NF-κB p65, p-IκB, and IκB in THP-1 or RAW264.7-derived M0, M2 and M2 macrophages pretreated with or without 740 Y-P prior to BJO treatment (1 µL/mL) were examined by western blot assay (n = 3). Data are expressed relative to the M0 group and presented as the mean ± SD of three independent experiments. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 vs. M0 group; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 vs. M2 group; ^&&p < 0.01, ^&&&p < 0.001 vs. M2 + BJO group

To determine whether BJO-mediated reprogramming of M2 macrophages depends on PI3K/AKT signaling, we used PI3K activator 740 Y-P. Co-treatment with 740 Y-P reversed BJO-induced suppression of AKT/mTOR phosphorylation and abolished the ability of BJO to repolarize M2 macrophages, as evidenced by the restoration of CD206 expression and reduction of CD86 to levels comparable to those in the M2 group (Fig. 5D). Furthermore, 740 Y-P substantially attenuated the anti-OS activity of M2-BJO-CM and reversed the BJO-induced increase in TNF-α and decrease in IL-10 secretion (Fig. S2C-F). We also found that inhibition of either ERK1/2 (by U0126) or NF-κB (by BAY 11-7082) was sufficient to block BJO-mediated M2-to-M1 repolarization, with CD206 and CD86 protein expression remaining similar to that of untreated M2 macrophages (Fig. 5E-F). Similarly, U0126 or BAY 11-7082 markedly impaired the inhibitory effect of M2-BJO-CM against OS and reversed its induction of TNF-α and suppression of IL-10 secretion (Fig. S2C-F). Given that the PI3K/AKT pathway negatively regulates MAPKs and NF-κB cascades in monocytes [26–27], we observed co-treatment with 740 Y-P reversed BJO-induced activation of these pathways, as evidenced by decreased phosphorylation of ERK1/2, NF-κB p65, and IκB-α, along with increased total IκB-α protein levels (Fig. 5G-H). Taken together, BJO induces repolarization of M2 macrophages by inhibiting the PI3K/AKT pathway, thereby relieving its negative regulation of ERK1/2 and NF-κB signaling.

BJO inhibits OS Progression and Reprograms TAMs In Vivo

To evaluate the anti-OS efficacy of BJO in vivo, mice were subcutaneously inoculated with K7M2 cells, either alone or in combination with IL-4-treated RAW264.7 cells, and then treated with or without BJO. Consistent with the in vitro findings, the presence of RAW264.7 cells markedly accelerated K7M2 tumor growth compared to K7M2 cells alone (Fig. 6A-B). This pro-tumoral effect was effectively inhibited by BJO, which suppressed tumor progression in both the K7M2-alone and co-injection models, reducing tumor volume by 53.69% and 87.92% (Fig. 6D) and tumor weight by 57.61% and 79.71% (Fig. 6C) at the endpoint, respectively, with greater suppression observed in the co-injection model.

Fig. 6.

Anti-OS efficacy of BJO in vivo. BALB/C nude mice subcutaneously inoculated with OS cells alone or co-inoculated with IL-4-treated RAW264.7 cells. When tumor volumes reached approximately 100 mm³, mice received daily oral gavage of vehicle (corn oil) or BJO (413 mg/kg) for 14 consecutive days. (A) Images of tumor-bearing mice and (B) harvested tumors after 14 days of treatment (n = 8). (C) Tumor weights, (D) tumor growth curve, and (E) body weight changes during the 14-day treatment period (n = 8). (F) Concentrations of NO, TNF-α, IL-10, and ARG1 in tumor tissues at endpoint (n = 8). (G) qPCR analysis of mRNA expression levels of Tnf, Nos2, Il1b, Il10, Mrc1, and Arg1 in tumor tissues at endpoint (n = 8). (H) Representative images of Ki67-, H&E- (scale bar = 50 μm), and CD31/PAS double-stained tumor sections (scale bar = 20 μm; black arrows indicate VM structures). Quantification of Ki67 expression and VM density is shown (n = 5). (I) Representative immunofluorescent images of F4/80 (green), CD86 (red), CD206 (purple) and DAPI (blue) in tumor sections (scale bar = 20 μm) and the percentages of CD86⁺ and CD206⁺ cells in the F4/80⁺ macrophages (n = 5). Data are expressed as the mean ± SD from at least five mice per group. ^**p < 0.01, ^***p < 0.001 vs. K7M2 group or K7M2 + M2 group; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 vs. K7M2 group; ^&p < 0.05, ^&&p < 0.01, ^&&&p < 0.001 vs. K7M2 + BJO group

In parallel, we evaluated the systemic tolerability of BJO in vivo. No significant body weight loss was observed through the treatment period (Fig. 6E). Histopathological examination of major organs revealed no apparent abnormalities (Fig. S3A). Moreover, serum levels of ALT, AST, BUN, and CRE were comparable to those in the blank control group (Fig. S3B), which collectively supported the safety of BJO at the tested doses.

To determine whether BJO exerts its anti-OS effects through TAM repolarization, we assessed macrophage polarization in vivo. ELISA analysis showed that tumors from the K7M2 + M2 co-injection group exhibited significantly higher levels of M2 markers (IL-10 and ARG1) compared to those from the K7M2-alone group. BJO treatment decreased levels of IL-10 and ARG1 levels while increasing TNF-α and NO production in OS tissues, with these effects being more pronounced in the co-injection model (Fig. 6F). Consistently, qPCR analysis confirmed these changes at the transcriptional level. Tumors from the co-injection group exhibited higher expression of M2-associated genes (Il10, Mrc1, Arg1) compared to the K7M2-alone group. BJO treatment upregulated M1-associated genes (Tnf, Nos2, Il1b) and downregulated M2-associated genes (l10, Mrc1, Arg1) in both models, again with more pronounced effects in the co-injection group (Fig. 6G). Immunofluorescence staining of OS tumor sections further supported these findings, showing that F4/80⁺CD206⁺ M2 macrophages were most abundant in tumors from co-injected mice. BJO treatment increased infiltration of F4/80⁺CD86⁺ M1 macrophages and reduced that of F4/80⁺CD206⁺ M2 macrophages in both models, with more pronounced changes in the co-injection model (Fig. 6H).

Histologically, the co-injection group exhibited higher Ki-67 positivity than those from the K7M2-alone group, consistent with their accelerated growth. BJO treatment markedly reduced the proportion of Ki-67–positive cells in the both OS models, with a more pronounced suppressive effect observed in the co-injection group. Moreover, BJO administration induced extensive tumor necrosis, characterized by eosinophilic cytoplasm and nuclear condensation with fragmentation. These pathological changes were most prominent in the co-injection group, further supporting the notion that BJO exerts stronger anti-OS effects within a TAM-enriched tumor milieu. To assess VM in vivo, we employed CD31-PAS dual staining on OS tissue sections. The K7M2 + M2 co-injection group exhibited the highest VM density among all groups, and BJO treatment reduced VM density by 70.83% in the K7M2 model and by 96.77% in the K7M2 + M2 model compared to their respective untreated controls (Fig. 6I). Taken together, these results suggest that BJO inhibits OS growth in vivo, at least in part, by reprogramming TAMs from an M2-like to an M1-like phenotype within the OS microenvironment.

Discussion

Over the past few decades, immunosuppression in the TME has become a major focus in cancer research. As the predominant immune cells infiltrating tumors, TAMs are often co-opted by cancer cells to act as accomplices, facilitating tumor progression, angiogenesis, and metastasis [28]. Given their critical role in promoting tumor progression, therapeutic approaches targeting M2-type TAMs within the TME have emerged as a promising strategy. Current efforts mainly concentrate on depleting M2-TAMs [7], inhibiting the CCL2/CCR2 axis responsible for their recruitment [29], or blocking M2-associated signaling pathways like CD47/SIRPα [30]. Herein, our study presents an alternative approach whereby Brucea javanica repolarizes M2-TAMs from a tumor-promoting M2 state toward an anti-tumor M1 phenotype.

Brucea javanica has long been recognized for its diverse pharmacological activities, including anti-tumor, anti-malarial, anti-diabetic and neuroprotective effects, with its potent anti-tumor properties being the most extensively studied [18]. Historically, these effects were attributed primarily to apoptosis or autophagy induction [18]; however, recent evidence shows that its main oil-based bioactive component, BJO, can remodel the TME in melanoma by shifting it from an immunosuppressive to an immunostimulatory state, as demonstrated by mass cytometry showing an increased M1-to-M2 ratio [19]. Among the constituents of BJO, palmitic acid inhibits gastric cancer progression by modulating macrophage polarization toward an M1-like phenotype [31]. Oleic acid, another active ingredient, is metabolized in vivo to oleamide, which exerts a potent anti-breast cancer effect by redirecting M2-like TAMs towards M1-like TAM phenotypes [32]. We therefore evaluated the ability of Brucea javanica to reprogram TAMs.

To this end, we first established classical M2 polarization models using human THP-1 and murine RAW264.7 macrophages stimulated with IL-4 and IL-13. In the TME, IL-4 and IL-13 promote the differentiation of macrophages into an M2-like phenotype, which exerts immunosuppressive effects and fosters a pro-tumorigenic microenvironment—thereby contributing to disease progression [12]. We next compared the repolarizing effects of BJO and DBA (the non-lipid fraction of Brucea javanica). At equivalent concentrations, BJO exhibited a more potent ability than DBA to repolarize M2 macrophages, as reflected by stronger induction of M1-associated genes (IL6, NOS2, and IL1B) and greater suppression of M2-associated genes (IL10, MRC1, ARG1, and CCL22), as well as a sharp increase in the CD86⁺ population and a more pronounced decrease in the CD206⁺ population compared to DBA. Moreover, the repolarizing effect of BJO was dose-dependent, characterized by a dose-responsive shift in the above-described M1- and M2-associated gene expression profiles at the mRNA level. Functionally, this was accompanied by enhanced production of canonical M1 effector molecules TNF-α and NO, along with concomitant downregulation of M2 markers Arg-1 and the anti-inflammatory cytokine IL-10, indicating that the oil fraction is the more active component in driving M2-to-M1 repolarization. Our GC-TOF-MS profiling further confirmed that BJO is rich in fatty acids such as oleic acid, linoleic acid, palmitic acid, and stearic acid. Consistent with their reported M1-polarization effects [31–32], these findings suggest that the fatty acid components may underlie the superior repolarizing activity of BJO compared with the non-lipid fraction.

These results prompted us to explore the functional consequences of M2 macrophage activity on OS cells and to determine whether BJO-induced reprogramming could counteract their pro-tumor influence. Consistent with earlier studies, we found that M2-CM markedly enhanced OS cell proliferation, invasion, and migration compared with M0-CM. In VM, tumor cells—rather than endothelial cells—form perfusable vascular channels through the acquisition of vasculogenic traits, representing an alternative form of angiogenesis frequently associated with advanced cancers such as OS [33–34]. To model VM in vitro, we performed a Matrigel tube formation assay using two OS cell lines previously reported to generate VM-like tubular structures [22]. As expected, treatment with M2-CM enhanced the ability of OS cells to form these structures, in line with previous evidence that TAMs facilitate VM across multiple malignancies [23]. Importantly, M2-BJO-CM lost the ability to promote malignant behaviors in OS cells, and BJO alone had no direct impact. It is reasonable to infer that the anti-OS effect of BJO is indirectly mediated through the reprogramming of M2 macrophages rather than through direct action on OS cells.

Given the established role of PI3K/AKT signaling in promoting M2 polarization [26] and its enrichment in the BJO–OS–TAMs target network, we hypothesized that BJO reprograms TAMs via this pathway to indirectly suppress OS progression. Prior evidence reported that BJO emulsion inhibits PI3K/AKT in Jurkat cells [35]. Supporting this notion, BJO suppressed IL-4-induced phosphorylation of PI3K, AKT, and mTOR in two M2 macrophage models. Importantly, 740 Y-P not only restored AKT/mTOR activation but also reversed BJO-mediated repolarization of M2 macrophages, eliminated the anti-OS activity of conditioned medium, and reverted cytokine secretion to a M2-like immunosuppressive profile—collectively indicating that PI3K/AKT inhibition is indispensable for BJO to reprogram TAMs from a pro-tumor to an anti-tumor phenotype. In parallel, BJO activated ERK1/2 and NF-κB—two signaling cascades essential for pro-inflammatory macrophage polarization [36]. Pharmacological blockade of either pathway was sufficient to abrogate both TAM repolarization and the consequent anti-OS effects, underscoring their essential contribution to BJO’s mechanism. Notably, studies in human PBMCs or THP-1 cells have reported that inhibition of the PI3K/AKT pathway by LY294002 or wortmannin can augment MAPK and NF-κB signaling—for instance, by relieving Akt-mediated suppression, thereby enhancing LPS-induced ERK1/2 phosphorylation, promoting NF-κB p65 nuclear translocation, and upregulating expression of downstream effectors such as TNF-α [37]. Further supporting this cross-talk, resveratrol, a natural compound, suppresses LPS-induced MAPK and NF-κB activation—and downstream inflammatory mediators such as iNOS and TNF-α—in BV-2 microglia, an effect reversed by the mTOR inhibitor rapamycin [38]. Although these observations were made in the context of LPS-activated macrophages, our data show that restoration of PI3K/AKT signaling by 740 Y-P attenuated BJO-induced ERK1/2 and NF-κB activation in M2 macrophages, confirming the role of PI3K/AKT in restraining inflammatory responses across macrophage phenotypes.

To study the role of TAMs in OS in vivo, we generated a syngeneic OS model by co-implanting K7M2 tumor cells with RAW264.7-derived M2-like macrophages. Our results demonstrated that M2 macrophages significantly enhanced OS tumor growth—a finding consistent with previous reports [39]—accompanied by increased Ki-67 positivity and higher VM density. Immunofluorescence, qPCR, and ELISA results collectively confirmed a pronounced M2-skewed phenotype in co-implanted tumors, demonstrating the suitability of this model for studying the relationship between TAMs and OS cells and assessing the therapeutic efficacy on OS through TAMs.

Previous studies showed that oral administration of 250 mg/kg BJO, typically quantified using oleic acid as the marker compound, results in plasma oleic acid levels of approximately 70 µg/mL in rats [40]. This exposure is comparable in magnitude to the BJO concentration expressed as oleic acid (~ 110 µg/mL) used in our in vitro experiments. Based on dose conversion factors, the rat dose of 250 mg/kg corresponds to roughly 360 mg/kg in mice; we therefore used 400 mg/kg BJO for in vivo studies. BJO demonstrated its greatest anti-OS efficacy in the TAM-enriched co-implantation model, accompanied by a shift in TAM polarization from M2- to M1-like phenotypes both in vivo and in our in vitro experiments, supporting that BJO acts primarily through reprogramming M2-TAMs. This enhanced efficacy in TAM-rich TME aligns with reports that some therapeutics show superior antitumor activity in M2 macrophage–co-injected models [41–42]. TAMs are thereby positioned not merely as bystanders but as central mediators of BJO’s anti-OS effect, underscoring TAM repolarization as a key mechanism underlying its efficacy in OS. Interestingly, appreciable tumor suppression was still observed in the K7M2-only model, albeit to a lesser extent. We speculate that this phenomenon is likely due to BJO’s ability to reprogram the endogenous M2-like TAMs inherently present in the TME of K7M2-only tumors [22]. The introduction of exogenous M2 macrophages in the co-implantation model amplified BJO’s anti-OS response by providing more cells for repolarization.

The in vivo safety of a therapeutic approach is crucial to its practical application. BJO treatment resulted in minimal changes in body weight and caused no pathological damage to major organs, indicating a favorable safety profile at the dose used in our study. This aligns with clinical evidence showing that BJO emulsion demonstrates a good safety and tolerability in cancer patients [43]. Hepatic and renal dysfunction are frequently observed in children with malignancies [44]. Consistent with this clinical observation, liver and kidney injury was detected in the OS-bearing model group. Notably, BJO treatment ameliorated these abnormalities. This protective effect is in line with previous reports demonstrating that oleic acid, the main ingredient of BJO according to the Chinese Pharmacopoeia, ameliorates experimental liver pathologies and attenuates reactive oxygen species-induced kidney injury [45–46].

Conclusion

In summary, we provide the first evidence that BJO attenuates OS progression through repolarizing TAMs from M2 to M1 phenotype. Mechanistically, BJO inhibits PI3K/AKT pathway in TAMs, thereby alleviating its negative regulation of ERK1/2 and NF-κB signaling. Our results implicate TAMs in the anti-OS effect of BJO and highlight its promise as a well-tolerated immunomodulatory agent for OS therapy.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Yongqi Guo: Data curation, Methodology, Writing – original draft; Yantao Jiang: Methodology, Investigation; Zhenlin Li: Formal analysis; Jing Zhou: Conceptualization, Investigation, Supervision; Nan Yao: Funding acquisition, Conceptualization, Investigation, Supervision; Writing – review and editing.

Funding

This work was supported by grants awarded by the National Natural Science Foundation of China (No. 82574637) and the Natural Science Foundation of Jiangsu Province (No. BK20231382).

Data Availability

The datasets presented in the study are available from the corresponding author on reasonable request.

Declarations

Clinical Trial Number

Not applicable.

Ethical Approval

Animal use was approved by the Ethics Committee of Jiangsu Province Academy of Traditional Chinese Medicine (Approval No. AEWC-20240202-368) and adhered to the established protocols for the ethical use and care of laboratory animals.

Consent to Participate

Not applicable.

Consent to Publish

Not applicable.

Competing interests

The authors have no relevant financial or non-financial interests to disclose.