Reprogramming tumor microenvironment via dual targeting co-delivery of regorafenib and alpha-difluoromethylornithine in osteosarcoma

Abstract

Background

Tumor angiogenesis, immunosuppression, and progression are all closely correlated with the tumor microenvironment (TME). Immune evasion is supported by both M2 phenotype tumor-associated macrophages (TAMs) and vascular aberrations in the TME. TME reprogramming is a promising therapeutic approach for treating tumors. Anti-angiogenesis has the power to control the polarization of macrophages, prevent progression, and increase drug penetration. Additionally, polyamine blocking therapy can increase CD8⁺ T cell infiltration and decrease immunosuppressive cells. These results led to developing a potential therapeutic regimen that targets TAMs and angiogenesis to reprogram the osteosarcoma TME.

Results

For the targeted biomimetic co-delivery of regorafenib and alpha-difluoromethylornithine via the mannose receptor, which is overexpressed in both TAMs and osteosarcoma cells, mannosylated poly(lactide-co-glycolide)-polyethylene glycol nanoparticles (Man-NPs) were synthesized. The superior physiological properties and intratumoral accumulation of the Man-NPs efficiently promoted TAMs polarization and inhibited angiogenesis. Macrophage repolarization further activated immune cells, which contributed to remodeling the TME.

Conclusion

Overall, these findings suggested that using Man-NPs as an immunotherapeutic approach to treat osteosarcoma may be promising.

Graphical Abstract

Background

The most common age groups for osteosarcoma (OS), a primary bone malignant tumor with mesenchymal tissue origin, are children and adolescents (Gill et al. 2021). The standard of care for patients with OS is surgery in combination with neo-adjuvant and adjuvant chemotherapeutic drugs, which appropriately provides 70% of patients a 5-year nonmetastatic survival rate. However, patients with lung metastasis had a five-year survival rate of less than 30% (Liu et al. 2021; Marchandet et al. 2021). Recently, immunotherapy has significantly changed how cancer is treated (Dallavalasa et al. 2021). The effectiveness of immunotherapy relies on tumor cells and the host immune system working together effectively. However, due to the low expression of PD-L1, osteosarcoma is regarded as a relatively “cold tumor”, demonstrating a poor response to immune checkpoint inhibitors PD-1 and PD-L1 (Galon et al. 2019; Tawbi et al. 2017). It is therefore widely accepted that altering the tumor immune microenvironment (TIME) is a potential strategy for triggering the body’s defenses against osteosarcoma.

As the majority of infiltrating immune cells in the tumor microenvironment (TME), tumor-associated macrophages (TAMs) have been used as a prognostic indicator and a new target for tumor immunotherapy (Koirala et al. 2016). M1 phenotype macrophages (termed TAM1) are thought to be anti-tumorigenic because they release cytotoxic chemicals and cytokines and promote T cell immunity (Cersosimo et al. 2020; Le et al. 2021). In contrast, M2 phenotype macrophages (termed TAM2) promote tumor angiogenesis and immune escape, ultimately leading to tumor progression and metastasis (Cersosimo et al. 2020; Le et al. 2021). However, tumor angiogenesis can also regulate the immune response by polarizing macrophages with the M2 phenotype and suppressing the immune system (Okikawa et al. 2022).

Inhibition of neovascularization by blocking the VEGF/VEGFR signaling pathway using multi-kinase inhibitors and anti-vascular endothelial growth factor (VEGF) monoclonal antibodies is a therapeutic strategy that influences immune cell populations in the TME. Regorafenib, a multi-kinase inhibitor with Food and Drug administration (FDA) approval, inhibits tumor progression by targeting the activity of oncogenic kinases that control cell signal transduction and tumor angiogenesis (Duffaud et al. 2019). More importantly, regorafenib could convert TAMs from M2 to M1 phenotype, induce T cell activation, and combine with anti-PD-1 sensitized tumors to produce synergistic immunomodulatory effects (Chiang et al. 2022; Doleschel et al. 2021; Ou et al. 2021). However, molecules that interfere with VEGF/VEGFR signaling are responsible for severe adverse effects because of inhibition of physiological angiogenesis (Lacal et al. 2018). Regorafenib significantly increased progression-free survival in phase II trials in patients with advanced osteosarcoma compared to placebo, but there was no appreciable statistical difference in overall survival (Davis et al. 2019). Here, a shift from a single- to a multi-target therapeutic approach is essential to trigger the host immune system and provide long-term efficacy.

Anti-angiogenic therapy has been proven to have some effect on TME, and recent literature suggests that additional drug therapies may also be able to modulate the immune system. Particularly, tumor progression in immunocompetent tumor-bearing mice was significantly suppressed by polyamine blocking therapy (PBT), which involved the inhibition of polyamine biosynthesis and transport (Hayes et al. 2014). During cellular processes, alpha-difluoromethylornithine (DFMO) is an FDA-approved inhibitor that inhibits ornithine decarboxylase and blocks polyamine biosynthesis (Pfeffer et al. 2000). In immunocompetent mice, DFMO therapy has been shown to inhibit tumor growth by increasing CD8⁺ T cell infiltration and reducing myeloid-derived suppressor cells activity (Ye et al. 2016). Furthermore, it has undergone favorable evaluation in clinical trials for the treatment and chemoprevention of neuroblastoma (Lewis et al. 2020; Samal et al. 2013; Sholler et al. 2018). Therefore, combining anti-angiogenic therapy with PBT may be a viable option to increase the effectiveness of immunotherapy, reduce the likelihood of drug side effects, and remodel the osteosarcoma TME.

In this study, we designed mannosylated poly(lactide-co-glycolide)-polyethylene glycol (PLGA-PEG) nanoparticles (Man-NPs) for biomimetic regorafenib and DFMO delivery that target the TME and remodel it. TAM2 and K7 osteosarcoma cells could specifically recognize and phagocytize the NPs due to the overexpression of CD206 on both types of cells. In vitro evaluations of Man-NPs’ physicochemical properties, regulation of TAM2 repolarization, and angiogenesis were thorough. To assess the in vivo tissue distribution and anti-tumor effectiveness of Man-NPs, we established a xenograft osteosarcoma model given the high degree of osteosarcoma malignancy. We discovered that Man-NPs accumulated more effectively in tumors, promoting the polarization of TAMs, the infiltration of CD8⁺ T cells, and the reduction of Tregs populations via anti-angiogenesis and PBT (Scheme 1) (created with BioRender.com). These findings offer a promising new strategy for TAMs-targeted immunotherapy and anti-angiogenesis that can effectively inhibit tumor growth by remodeling the TME.

Scheme 1

Schematic representation of tumor immunotherapy using Man-NPs, which regulates tumor-associated macrophages and angiogenesis in tandem to target the tumor microenvironment

Results and discussion

Properties and characterization of Man-NPs

The mannose receptor, also named as CD206, is one of the most frequently used TAMs-targeted receptors, owing to its high expression in TAM2. As shown in Fig. 1a, TAM2 and, importantly, K7 cells both expressed CD206 at high levels. As a result, mannose-modified NPs that target TAMs and osteosarcoma cells can be designed rationally because they have a physiological bias. In the current study, mannose was conjugated with PLGA-PEG as a targeted ligand of TAM2 and osteosarcoma cells based on the specific identification of mannose receptors. The H¹NMR spectrum was used to determine the man-PLGA-PEG co-polymer (Additional file 1: Fig. S1). A double emulsion solvent evaporation method successfully created the mannosylated nanoparticles enclosed in regorafenib/DFMO. Man-NPs had an average particle size of 146 nm, while unmodified PLGA-PEG nanoparticles, or NPs, had a particle size of 125 nm (Fig. 2b, c). The partial neutralization of the negative charge by mannose caused the zeta potentials of NPs to change after the mannose modification from − 23.3 mV to − 10.6 mV (Additional file 1: Fig. S2a). Man-NPs and NPs exhibited low PDI values (< 0.3), indicating narrow distribution of molecular weight (Additional file 1: Fig. S2b). The drug encapsulation efficiency of Man-NPs was 77.68% for DFMO and 74.6% for regorafenib, and the drug-loading efficiency was 2.9% and 0.6%, respectively. Man-NPs also displayed a sustained-release pattern (Fig. 1d). In 10% FBS, the Man-NPs displayed good stability (Additional file 1: Fig. S2c), which is appropriate for their biomedical applications.

Fig. 1

In vitro characterization of Man-NPs. a CD206 expression in K7 cells, K7 tumor, M1 and M2 phenotype macrophages. Size distribution and TEM image of the NPs (b) and Man-NPs (c). d Cumulative release of regorafenib in Man-NPs. Uptake efficiency of NPs and Man-NPs with or without pretreatment with mannose in K7 cells (e), TAM2 (f), and TAM1 (g). Representative CLSM images of K7 cells (h) and M2 phenotype macrophage (i) after 6 h of incubation with the Coumarin-6-loaded NPs (scale bar: 25 μm). Comparisons were made using the Student’s t-test; ns: no significant difference, **p < 0.01, ***p < 0.001

Fig. 2

In vitro anti-tumor effect of Man-NPs. a Cytotoxicity study for K7 cells treated with drugs for 48 h. b Apoptosis assay of K7 cells via FACS. The right panel displayed the results of the apoptosis rate. c Western blotting was used to detect apoptotic protein. d VEGFA level of K7 cells after treatment. e Schematic of the transwell co-culture model. f FACS assay for measuring apoptosis in K7 cells co-cultured with TAM2. The right panel displayed the results of the apoptosis rate. g Western blotting was used to identify the apoptotic protein produced by K7 cells co-cultured with TAM2. h VEGFA levels in K7 cells co-cultured with TAM2. Comparisons were made using one-way ANOVA with Tukey’s test; ns: no significant difference, *p < 0.05, **p < 0.01, ***p < 0.001

To detect whether Man-NPs were preferentially absorbed by K7 cells and TAM2, we first polarized BMDMs into TAM1 or TAM2, and then incubated them with coumarin-6-loaded Man-NPs and NPs for 6 h. The mean fluorescence intensity of the Man-NPs absorbed by the K7 cells and TAM2 in the flow cytometry assay was higher than that of the unmodified NPs. The cellular uptake efficiency of the Man-NPs was significantly lower when the cells were pre-treated with free mannose. Additionally, due to the low expression of CD206, TAM1 almost equally absorbed Man-NPs and NPs (Fig. 1e–g; Additional file 1: Fig. S2d). The CLSM assay results were consistent with flow cytometry (Fig. 1h, i), confirming the theory that Man-NPs are more readily taken up by cells when the mannose receptor mediates them. These findings suggested that the mannose receptor is a key target in mediating drug delivery.

In vitro cytotoxicity

The therapeutic effect should be enhanced owing to the increased cellular uptake of Man-NPs by K7 cells. To validate this, we assessed the anticancer activity of Man-NPs by CCK-8 assay. Compared to free regorafenib, DR, or unmodified NPs, Man-NPs significantly inhibited the proliferation of the K7 cells in a dose- and time-dependent manner and had the lowest IC₅₀ value (Fig. 2a; Additional file 1: Fig. S3a-b). However, approximately 90% of TAM1 and TAM2 did not experience any side effects and tolerated the Man-NPs and NPs treatment well (Additional file 1: Fig. S3c). Regorafenib inhibits cancer cell growth by inducing apoptosis and deactivating the AKT and ERK signaling pathways (Pan et al. 2019; Weng et al. 2019). Through the suppression of the MAPK/ERK and AKT/mTOR signaling pathways, DFMO slows the progression of esophageal cancer (Xie et al. 2020). Man-NPs demonstrated strong anti-tumor efficacy in this study by significantly suppressing the expression of p-AKT, p-MAPK, p-ERK, and p-mTOR (Additional file 1: Fig. S3d), known as signaling pathways for cell survival and growth. Furthermore, Man-NPs were used to test K7 cells for cell apoptosis. In comparison to other treatments, Man-NPs caused a higher rate of apoptosis in K7 cells, as shown in Fig. 2b. Additionally, following Man-NPs treatment, the expression of cleaved-caspase3 was increased. The anti-apoptotic proteins Bcl-xl and Bcl-2 were significantly downregulated (Fig. 2c). One of the crucial cytokines in TME is VEGF. VEGFA, a member of the VEGF family primarily released by cancer cells and TAMs, is essential for tumor angiogenesis (Valkovic et al. 2002). Meanwhile, VEGFA is an immunosuppressive agent that can suppress T cell activity while inducing the accumulation of Tregs (Bourhis et al. 2021). Compared to unmodified NPs, the secretion of VEGFA significantly decreased after Man-NPs treatment (Fig. 2d).

We used a transwell co-culture model to examine how TAMs affect tumor cells while accurately simulating the physiological conditions of all cells exposed to the drug in TME. In the transwell co-culture system of K7 cells and TAM2 (Fig. 2e), Man-NPs also increased the rate of apoptosis (Fig. 2f) and the expression of apoptosis-related proteins (Fig. 2g), which reduced VEGFA secretion compared to other groups (Fig. 2h).

In vitro polarization and repolarization of macrophages

TAMs are highly dynamic and malleable with reversible polarization between M1 and M2 phenotypes, and they have a variety of roles that include involvement in tumor growth, metastasis, angiogenesis, and immunosuppression (Dallavalasa et al. 2021). Numerous studies have shown that lung metastasis and a poor prognosis in patients with osteosarcoma were significantly correlated with a high density of TAM2 infiltration (Song et al. 2020; Yang et al. 2021). Therefore, reprogramming TAMs is a potential method for treating osteosarcoma without immunosuppression. Figure 3a showed a schematic representation of how BMDMs are converted into macrophages with the M1 and M2 phenotypes. The phenotypes of the macrophages were analyzed by western blot. Man-NPs treatment significantly reduced CD206 expression, suggesting that Man-NPs were successful in reversing TAM2 polarization (Fig. 3b). Man-NPs reversed the polarization toward TAM2 in the transwell co-culture system of TAMs and K7 cells (Fig. 3c), as demonstrated by the downregulation of CD206 (Fig. 3d). The mRNA expression of macrophage biomarkers, including CD86, TNF-α, CD206, and TGF-β, were detected using qRT-PCR to validate this finding further. As shown in Fig. 3e, f, Man-NPs significantly reduced the expression of CD206 and TGF-β (M2 markers) in both TAM1 and TAM2 while substantially increasing the expression of CD86 and TNF-α (M1 marker) in comparison to the untreated cells. IL-10 produced by TAMs can reduce the anti-tumor effect by inhibiting the antigen-presenting cells and reducing T cell effector capabilities like cytotoxicity (Ouyang et al. 2019). The Man-NPs treatment significantly decreased the levels of pro-tumor cytokine IL-10 (Fig. 3g), a stimulator of STAT3 (Verma et al. 2012). Inhibiting the STAT3 signaling pathway can cause macrophage phenotype to change from M2 to M1, as phosphorylated STAT3 is an essential step in TAM2 polarization (Solis-Martinez et al. 2018). Our findings showed that Man-NPs inhibited STAT3 phosphorylation in TAM1 and TAM2, which was in accordance with a prior study of the effect of DFMO on macrophage polarization through the STAT3 pathway (Alexander et al. 2020). Man-NPs in the TAMs/K7 transwell co-culture system increased pro-inflammatory NF-κB, essential for immune and inflammatory responses, through STAT3 inhibition (Fig. 3h, i). Together, our data suggested that Man-NPs strongly encourage TAM2 repolarization and have the potential to remodel TIME.

Fig. 3

In vitro control of macrophage repolarization. a Schematic illustration of BMDMs isolation, culture, and stimulation. b CD206 expression were detected by western blotting in TAMs after drug treatment. c Schematic of the transwell co-culture model. d The expression of CD206 in the drug-treated TAM2 co-cultured with K7 cells. The mRNA levels of CD86, TNF-α, CD206, and TGF-β in the drug-treated TAM1 (e) and TAM2 (f) co-cultured with K7 cells. g ELISA analysis of the amount of secreted IL-10 in the co-culture of K7 cells and TAM2 after drug treatment. Expression of p-STAT3 and NF-κB in TAM1 (h) and TAM2 (i) co-cultured with K7 cells after drug treatment. Comparisons were made using one-way ANOVA with Tukey’s test; ns: no significant difference, *p < 0.05, **p < 0.01, ***p < 0.001

In vitro anti-angiogenesis by Man-NPs

An important proof called tumor angiogenesis is what gives rapidly expanding malignant tumors their essential nutrients and oxygen. Several studies have demonstrated that TAMs, particularly TAM2, are crucial for developing abnormal blood vessels in tumors because they increase VEGF release (Riabov et al. 2014). During the development of tumors, VEGF interacts with the VEGF receptor (VEGFR) on endothelial cells to promote endothelial cell proliferation and migration and angiogenesis close to the tumor site (Dallavalasa et al. 2021). Thus, in the transwell co-culture system of TAMs and HUVEC (Fig. 4a), anti-angiogenesis was investigated using the Matrigel tube formation assay. According to Fig. 4b, TAM2 was primarily responsible for producing VEGF secretion, which was significantly decreased in co-cultured TAM2 cells with HUVEC after Man-NPs treatment. In the capillary-like network, TAM1 inhibited tube formation while TAM2 encouraged it. Man-NPs blocked this encouraged effect as a result of TAM2 repolarization toward TAM1 (Fig. 4c). Furthermore, Man-NPs inhibited the expression of VEGFR2 and p-VEGFR2 in HUVEC (Fig. 4d), which demonstrated a potent anti-angiogenesis effect. These results revealed that Man-NPs inhibit VEGFA/VEGFR2 signaling and regulate macrophage polarization to inhibit tumor angiogenesis.

Fig. 4

Anti-angiogenic effect of Man-NPs. a Schematic of a transwell co-culture model. b The level of VEGFA in the drug-treated TAM1 or TAM2 co-cultured with HUVEC. c Tube formation assay in the drug-treated TAM1 or TAM2 co-cultured with HUVEC (scale bar: 100 µm). d Expression of VEGFR2 and p-VEGFR2 in HUVEC was detected by Western blotting. Comparisons were made using one-way ANOVA with Tukey’s test; ns: no significant difference, *p < 0.05, **p < 0.01

In vivo tissue biodistribution

The tissue biodistribution of IR780-labeled nanoparticles were investigated in the K7 osteosarcoma tumor-bearing model to evaluate the in vivo tumor-targeting capability of Man-NPs. As shown in Fig. 5a–e, the IR780-labeled NPs and the Man-NPs were successfully delivered to the tumor site. Still, at all times studied, the Man-NPs-treated mice displayed a superior tumor accumulation to the unmodified NPs-treated mice. Ex vivo fluorescent images of the excised tumors further demonstrated that Man-NPs-treated groups had higher tumor accumulation than NPs-treated groups, suggesting that the targeting function provided by the mannose receptor is beneficial.

Fig. 5

In vivo imaging and drug-targeting delivery of Man-NPs. a Fluorescence imaging was used to detect drug accumulation in the tumors. b Radiant efficiency in vivo at the tumor sites. c Ex vivo fluorescence imaging of tumors. d Ex vivo radiant efficiency of tumors. e Ex vivo fluorescence imaging of major organs. f The intratumoral distribution of the IR780-loaded Man-NPs with staining of CD206 (scale bar: 25 µm). Comparisons were made using the Student’s t-test; **p < 0.01

Furthermore, the colocalization of CD206 and Man-NPs was performed to measure intratumoral penetration using immunofluorescent staining. The intratumoral distribution of IR780-labeled Man-NPs was largely consistent with the expression of CD206, as demonstrated by the immunofluorescence imaging (Fig. 5f). According to all of the findings above, the ability of Man-NPs to target tumors is due to the relatively higher phagocytosis of TAMs and cancer cells based on the mannose receptor.

In vivo anti-tumor efficacy and safety evaluation

The murine K7 osteosarcoma subcutaneous tumor model was used to calculate the impact of Man-NPs on tumor growth inhibition in vivo. Anti-tumor therapy was started on the eighth day following subcutaneous injection, and the drugs were injected intravenously every other day for approximately two weeks, as shown in Fig. 6a. The rates of tumor growth inhibition in the Regorafenib and DFMO groups were 39.0% and 22.7%, respectively, while the DR group increased the rate to 56.8%. The strongest anti-tumor effect was seen in Man-NPs group, with a tumor inhibition rate of 69.8% compared to 59.9% in the NPs group. Furthermore, the Man-NPs group in tumor-bearing mice had the longest survival time (Fig. 6b–k). The lowest cell proliferation capacity was observed by Ki67 staining assays following Man-NPs treatment, which was consistent with tumor growth inhibition (Fig. 6l). Our findings demonstrated the value of targeted delivery for enhancing anti-cancer effectiveness.

Fig. 6

In vivo therapeutic efficacy of Man-NPs. a A schematic illustration of the treatment schedule. b The tumor growth curve in K7 subcutaneous tumor-bearing mice following drug treatment (n = 6). (c–h) Individual tumor growth curves in different groups (n = 6). Tumor weight (i), tumor growth inhibition ratio (j) and survival curve (k) after drug treatment (n = 6). l Immunohistochemical staining (brown color) of Ki-67 (scale bar: 50 µm). The right panel displayed the percentage of Ki67 positive results. Comparisons were made using one-way ANOVA with Tukey’s test; ns: no significant difference, *p < 0.05, **p < 0.01, ***p < 0.001

Additionally, the preliminary biosafety of Man-NPs was evaluated. The main organs (heart, liver, spleen, lung, and kidney) did not significantly change pathologically during treatment, nor was there any obvious loss of body weight (Additional file 1: Fig. S4a, b). These results indicated that Man-NPs are nontoxic to these organs and exhibit strong biosafety.

In vivo remodeling TIME and inhibiting angiogenesis

The in vivo repolarization of TAMs was assessed by measuring the proportion of M1/M2 phenotype, cytokine secretion, and the expression profile of biomarker proteins to investigate the promising immunotherapeutic effects of Man-NPs. A flow cytometry assay detected the proportion of TAMs of various phenotypes within the tumor tissues. Intratumor TAMs in the PBS group primarily comprised of M2 phenotype (F4/80⁺CD206⁺) while M1 phenotype macrophages (F4/80⁺CD86⁺) were rarely expressed, which may have contributed to the rapid tumor growth. Regorafenib and DFMO, either alone or in combination, increased the proportion of M1 phenotype macrophages and decreased the proportion of M2 phenotype macrophages. At the same time, the Man-NPs treatment group showed more noticeable changes (Fig. 7a, b; Additional file 1: Fig. S5a, b). The proportion of M1/M2, which is related to the polarization effect (Zhao et al. 2021), was analyzed to more accurately estimate the in vivo polarization capacity of Man-NPs. As shown in Fig. 7c, the population of total TAMs remained fairly constant among the various groups. Still, the ratio of M1(CD86⁺)/M2 (CD206⁺) of the Man-NPs group was remarkably higher than that of other treatment groups, suggesting that Man-NPs targeted delivery of nanomedicine may have effectively in vivo reprogrammed macrophages from M2 to M1 phenotype.

Fig. 7

In vivo remodeling TME. The percentage of TAMs with CD206⁺ (a) and CD86⁺ cells (b) was determined by FACS in subcutaneous tumor tissues. c The intratumoral ratio of CD86⁺ M1/CD206 + M2 phenotype. The percentage of CD8⁺ cells (d) and granzyme B⁺ CD8⁺ T cells (e) in intratumoral infiltration. The percentage of CD4⁺ cells (f) and Foxp3⁺ CD4⁺ T cells (g) in intratumoral infiltration. ELISA analysis of intratumoral cytokine secretion of IFN-γ (h) and TNF-α (i). j The expression of CD206, HIF-1α and VEGFR2 were detected by western blotting in subcutaneous tumor tissues after drug treatment. k Immunohistochemical staining (brown color) of CD86, CD86, and CD31 (scale bars: 100 µm). Comparisons were made using one-way ANOVA with Tukey’s test; ns: no significant difference, *p < 0.05, **p < 0.01, ***p < 0.001

It is known that TAM2 can suppress the immune system by preventing tumor-infiltrating lymphocytes from activating (Kuroda et al. 2021). We also conducted a quantitative analysis of CD4⁺ helper T cells and CD8⁺ cytotoxic T cells to evaluate the activation of the anti-tumor immune reaction following various treatments. As shown in Fig. 7d–g and Additional file 1: Fig. S5c, d, the Man-NPs treatment significantly increased the tumor population of cytotoxic CD8⁺ granzyme B⁺ T lymphocytes to 29.4%, compared to 6.9% in the PBS group. In contrast, the population of CD4⁺ T cells barely changed among the various groups. Notably, compared to the other groups, there was a significantly lower percentage of intratumor infiltrating CD4⁺ Foxp3⁺ T cells (Tregs), which are associated with immune evasion and a poor prognosis in patients with cancer (Dannenmann et al. 2013). Additionally, following Man-NPs treatment, elevated CD8⁺ T lymphocytes and declined CD4⁺ T cells in the spleen were detected (Additional file 1: Fig. S5e). TAM1 and CD8⁺ T cells produce immunogenic cytokines like IFN-γ and TNF-α to improve the immune response against tumors (Le et al. 2021). Therefore, ELISA was used to determine the intratumoral cytokine levels. The levels of anti-tumor IFN-γ and TNF-α were increased after treatment with Man-NPs due to the enhancement of TAM1 and CD8⁺ T cells in the tumor tissues (Fig. 7h, i). These findings showed that macrophage repolarization increased immune response activation, which helped to remodel the TIME.

In addition to analyzing the ratio of M1/M2 macrophages and cytokine secretion, western blotting and immunohistochemical staining were used to determine the expression profiles of macrophage surface proteins (CD206 for M2 marker and CD86 for M1 marker) and angiogenic marker proteins (VEGFR2, HIF-1α, and CD31) to further confirm the polarization ability and anti-angiogenesis of Man-NPs. As shown in Fig. 7j, k, the results of the proportion of M1/M2 macrophages were consistent with the intratumoral CD206 expression, which was downregulated, while CD86 expression was upregulated after Man-NPs treatment. The expression of the angiogenic proteins HIF-1α and VEGFR2 was downregulated, which impacted the tumor vessel density (CD31 staining). According to all the findings above, Man-NPs could reprogram TME by boosting TAMs polarization, immune cell infiltration, and angiogenesis inhibition to improve anti-tumor activity in vivo.

Conclusions

In this study, we designed a combination therapy for osteosarcoma that targets TME by dual modulation on TAMs and angiogenesis using a mannosylated PLGA-PEG biomimetic system containing regorafenib and DFMO. By encouraging macrophage polarization from the pro-tumorigenic M2 to the anti-tumorigenic M1 phenotype and inhibiting angiogenesis, the prepared Man-NPs could produce a synergic anti-tumor efficacy with negligible systemic toxicity. Additionally, macrophage repolarization increased immune effector cell activation (e.g., increased CD8⁺ T cell infiltration and decreased Tregs), which helped to remodel the osteosarcoma TME. These excellent results demonstrated the effectiveness of Man-NPs as a targeted, safe, and powerful anti-tumor nanomedicine, suggesting that it may be a promising immunotherapeutic approach for treating osteosarcoma, with strong development and translational potential.

Materials and methods

Materials

Regorafenib, alpha-difluoromethylornithine, and Coumarin-6 were purchased from MedChemExpress (Monmouth Junction, NJ, USA). Shandong Academy of Pharmaceutical Sciences (China) provided PLGA_10K-PEG_5K-NH₂. Poly(vinyl alcohol) [PVA, 87–89% hydrolyzed, molecular weight (Mw) 31,000–50,000], lipopolysaccharide (LPS), 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride (EDC), and IR 780 iodide were purchased from Sigma Aldrich (St. Louis, MO, USA). Bide Pharmatech Ltd. (Shanghai, China) provided the A-d-mannosylphenyl isothiocyanate. We purchased sulfo-n-hydroxysuccinimide (NHS) from Aladdin Holdings Group Co., Ltd. (Beijing, China). Murine IL-4 and M-CSF were supplied by PeproTech (Rocky Hill, NJ, USA).

Mannose receptor (CD206), CD31, and β-Actin were received from Abcam (UK). Bcl-2, Bcl-xl, cleaved-caspase3, p-STAT3, NF-κB, VEGFR2, p-VEGFR2, p-MAPK, p-ERK, p-mTOR, Ki-67, and GAPDH were provided with Cell Signaling Technology (USA). R&D Systems (USA) provided fixable viability stain 620 and APC-Cy7 anti-mouse CD45. The following antibodies were obtained from BioLegend (USA): FITC anti-mouse/human CD11b, PerCP-Cy5.5 anti-mouse F4/80, Pacific Blue™ anti-mouse CD86, PE/Cy7 anti-mouse CD206, FITC anti-mouse CD45, APC- Fire™ 750 anti-mouse CD3, PE-Cy7 anti-mouse CD4, PerCP-Cy5.5 anti-mouse CD8 alpha, PE anti-mouse CD25, Alexa Fluor^® 647 anti-mouse Foxp3, and AF700 anti-mouse Granzyme B. MultiSciences Biotech Co., Ltd (Hangzhou, China) provided an enzyme-linked immunosorbent assay (ELISA) kit for the detection of mouse VEGFA, IL-10, tumor necrosis factor-alpha (TNF-α), and interferon-gamma (IFN-γ). Sinapharm Chemical Reagent Co., Ltd. (Shanghai, China) provided all other substances.

Synthesis of the drug-encapsulated nanoparticles

A prior study used a double emulsion solvent evaporation method to create the drug-encapsulated NPs (Zambito et al. 2021). Briefly, 3 ml of dichloromethane (DCM) containing 300 mg PLGA-PEG (organic phase) were combined with 5 mg DFMO solution (100 μL deionized water) as the first aqueous phase. Then, while stirring, 100 μL of DMSO containing 1 mg of regorafenib was added to the polymer-DCM solution. Using ultrasound on ice for 30 s, the primary water in oil (w/o) emulsion was obtained. The primary emulsion was then added to 15 mL of 1.0% (w/v) PVA water solution and homogenized using an ultrasonic device in an ice bath for 30 min. The water–oil-water (w/o/w) emulsion was then added to 50 mL of 0.3% PVA and stirred slowly overnight at room temperature (RT) to sanction the solvent to evaporate and nanoparticles to solidify. Finally, NPs were lyophilized, stored at − 20 °C until further use, and rinsed with deionized water three times. Similar techniques were used to create NPs loaded with coumarin-6 or IR780.

Preparation and characterization of Man-NPs

A-D-mannosylphenyl isothiocyanate was attached to the surface of drug-encapsulated NPs to create Man-NPs. Briefly, 10 mg of NHS and 10 mg of EDC were added to 100 mg of NPs dissolved in 10 mL MES buffer (pH 5.0) and stirred for 2 h. The mixture solution was given 100 mg of A-d-mannosylphenyl isothiocyanate and was stirred overnight at RT. Following a 12 h period, Man-NPs were precipitated with cold diethyl ether and methanol and dried under a vacuum. The co-polymer was precipitated, washed with cold methanol, and dried under a vacuum to remove unreacted mannose and excess reagents. H¹NMR spectrum of synthesized Man-PLGA-PEG was used to identify its chemical composition in CDCl3.

The dynamic light scattering measurement (Malvern, UK) was used to determine the particle size, zeta potential, and polydispersity index (PDI) of the NPs. The morphology of the NPs was observed using transmission electron microscopy (H-7650; Hitachi, Japan). The encapsulation efficiency and drug-loading capacity of regorafenib in NPs were determined by high-performance liquid chromatography (HPLC) equipped with Agilent-C18 (4.6 × 250 mm, 5 μm) column (1260 Infinity, Agilent Technologies, USA) with a flow rate of 0.5 mL/min and column temperature of 35 °C. The mobile phase contained 0.5% KH2PO4 (pH 3.5) and 70% acetonitrile, and the detection wavelength was 260 nm. By using liquid chromatography-mass spectrometry (LCMS) and an Acquity UPLC BEH C18 (2.1 × 100 mm, 1.7 μm) column (ACQUITY QDa, Waters Corp., USA) with a flow rate of 0.3 mL/min and column temperature of 35 °C, the encapsulation efficiency and drug-loading capacity of DFMO in NPs were detected. 100% acetonitrile and 0.1% formic acid aqueous solution made up the mobile phase. The calculation was as follows:

$${\text{Encapsulation efficiency }}\left( \% \right) \, = \, \left( {\text{mass of encapsulated drug}} \right)/\left( {\text{mass of drug added}} \right) \, \times \, 100\%$$

$${\text{Drug-loading capacity }}\left( \% \right) \, = \, \left( {\text{mass of encapsulated drug}} \right)/\left( {\text{mass of NPs}} \right) \, \times \, 100\%$$

Stability of NPs and in vitro drug release

The NPs were dissolved in phosphate buffer saline (PBS) containing 10% FBS to assess the stability. At various time points, the NPs changing particle size was examined. Cumulative in vitro drug release of Man-NPs was evaluated using a dialysis method. Man-NPs were placed into a dialysis bag (MWCO, 8–14 kDa), suspended in a mixed solution of PBS with 0.05% SDS and 60% methanol at two different pH values (7.4 and 6.5), and then gently shaken at 100 rpm at 37 °C. A corresponding amount of fresh medium was added, and 1 mL of the release medium was taken at the specified time to measure the concentration of regorafenib by HPLC.

Cell culture

American Type Culture Collection (Manassas, VA, USA) provided the murine osteosarcoma cell line K7, RAW264.7 and human umbilical vein endothelial cells (HUVEC). K7 cells and RAW264.7 were cultured in DMEM containing 10% FBS plus 1% penicillin/streptomycin (Gibco, USA). HUVEC were cultured in the endothelial cell medium (ECM) supplemented with 5% FBS, and 1% endothelial cell growth supplement (ECGS) (ScienCell, USA). All cell lines were maintained in a 37 °C incubator with 5% carbon dioxide and 95% air.

Collection and polarization of bone marrow-derived macrophages (BMDMs)

As previously described (Pineda-Torra et al. 2015), C57BL/6 mice’s femurs and tibias were used to collect bone marrow cells. The cells were cultured in RPMI 1640 medium with 10% FBS and 20 ng/mL M-CSF for 5 days to promote differentiation into BMDMs. Following differentiation, BMDMs were stimulated for 24 h with either 100 ng/mL LPS or 40 ng/mL IL-4 to polarize them into the M1 or M2 phenotypes, respectively. The macrophage phenotypic marker was further detected using real-time quantitative polymerase chain reaction (qRT-PCR), western blotting, and ELISA.

In vitro cellular uptake study

The K7 cells were cultured in a 6-well plate and 35-mm confocal dishes for 24 h with 2 × 10⁵ cells/well or 1 × 10⁵ cells/dish. The coumarin-6-loaded NPs and Man-NPs were added to the cells and incubated for 6 h. Coumarin-6 was quantified by fluorescence spectrophotometer (Hitachi, Japan) with an excitation wavelength of 420 nm and an emission wavelength of 506 nm. The cells were pre-treated with mannose (100 mM) for 4 h in a parallel group before incubating with the Man-NPs (Su et al. 2015). The cellular uptake efficiency was then measured using a flow cytometer assay (BD Biosciences, USA). Additionally, the cells in confocal dishes underwent three PBS washes, a 15 min fixation in 4% paraformaldehyde and DAPI staining, and were observed under a confocal laser scanning microscope (CLSM; TCS-SP8, Leica, Germany). Similarly, the cellular uptake of stimulated macrophages was carried out as described above.

In vitro cytotoxicity study

The K7 cells were plated in the 96-well plates at a density of 4000 cells per well and cultured for 24 h. Free regorafenib, DFMO, combined DFMO and regorafenib (DR), NPs, or Man-NPs, were all used to treat the cells. According to the manufacturer’s instructions, cell counting kit-8 (CCK-8; Dojindo) were used to test the viability of the cells after 24 h, 48 h and 72 h of incubation.

Preparation of transwell co-culture model

A transwell system with a 0.4 μm pore membrane (Corning, USA) was used to create a co-culture model. Its purpose was to faithfully reproduce the physiological conditions of TME, where the cells were exposed to the drugs. The transwell system's upper and lower compartments were seeded with TAM1 or TAM2 and K7 cells and co-incubated for 12 h. All cells were exposed to free regorafenib (2 μM), DFMO (1 mM), DR, NPs, or Man-NPs for 24 h. The K7 cells were harvested for further experiments. However, If TAM1 or TAM2 were to be used for analysis, K7 cells would be seeded in the upper compartment to facilitate the collection process.

Similarly, TAM1 or TAM2 was seeded into the transwell system's upper compartment, while HUVECs (2 × 10⁴) cells were cultured in the lower compartment after being treated with a matrigel matrix. Following the above-mentioned treatments, the cells were incubated at 37 °C for 12 h. The capillary network was photographed using the fluorescence microscope (Leica, Germany).

Flow cytometry analysis

The cells were collected after being co-incubated for 24 h, washed, and further stained by an Annexin V-FITC/PI apoptosis detection kit (Thermo Fisher, Waltham, USA) according to the manufacturer's instructions to quantitatively analyze the K7 cell apoptosis rate after drug treatment with or without co-culture. Furthermore, flow cytometry was used to quantitatively analyze the proportion of TAMs and T cells infiltrating tumor and spleen to assess the in vivo polarization ability of Man-NPs. Red blood cells were then separated from the fresh tumor and spleens tissues using red blood cell lysis buffer after the tissues had been cut into slices and digested with hyaluronidase, type IV collagenase, and DNA enzyme. The cells were extracted using lymphocyte separation solution for T cell sorting. Then they were incubated with specific immune cell biomarkers like CD86, CD206, Foxp3, and Granzyme B. After being filtered, the cell suspensions were examined with the FACS Calibur.

Western blot

After treatment, the cells and tumor tissues were harvested and lysed for 30 min in an ice bath using RIPA buffer mixed with a protease and phosphatase inhibitor (Sigma Aldrich, USA). The supernatants were collected after centrifugation at 12,000g at 4 °C for 20 min, and the results were determined using the BCA assay (Beyotime, China). The protein samples were separated by SDA-PAGE and then transferred to PVDF membranes (Millipore, USA). After incubating TBST containing 5% non-fat milk for 1 h at RT, these membranes were incubated overnight at 4 °C with primary antibodies. Following a standard procedure, secondary antibodies conjugated to horseradish peroxidase were incubated with the protein bands. Finally, signals were visualized using ECL detection reagents, and the loading control was either GAPDH or β-actin.

qRT-PCR and ELISA

The total RNA was extracted from the macrophages with TRIzol Reagent (Thermo, USA) and was reverse-transcribed into cDNA using the TaKaRa PrimeScript RT Reagent kit (TaKaRa, Japan) according to the manufacturer’s instructions. qRT-PCR was then performed using SYBR^® Premix DimerEraser^™ (TaKaRa, Japan) following the manufacturer's instructions. mRNA levels were normalized using β-actin as a control. The gene-specific primer sequences are listed in Additional file 1: Table S1. Using ELISA kits following the manufacturer's instructions, the levels of cytokines (VEGFA, IL-10, IFN-γ, and TNF-α) in the culture medium or tumor tissues were determined.

In vivo imaging and biodistribution study

Through the subcutaneous injection of K7 cells (1 × 10⁶ cells) into the back of BALB/c mice (female, 5–6 weeks old), the xenograft osteosarcoma tumor model was created. The mice were randomly divided into two groups when the tumors reached a median size of 500 mm³. Each mouse received an equal dose of IR780-loaded NPs or Man-NPs injected into the tail vein. Using an IVIS imaging system (Xenogen), the biodistribution of the NPs in mice was studied at various time points. The mice were euthanized after 24 h, and the tumors and major organs were surgically removed for ex vivo imaging. The tissue from the K7 osteosarcoma tumor model was also obtained for cryosection to conduct the co-localization study. Slices were stained with Alexa Fluor 488-conjugated goat secondary antibody for 1 h at RT, followed by DAPI staining. The slices were first incubated with an anti-CD206 antibody at 4 °C overnight. The co-localization was observed using CLSM.

In vivo anti-tumor efficacy and safety evaluation

As mentioned above, the osteosarcoma tumor model was developed in BALB/c mice. For the administration of drugs, the tumor-bearing mice were randomly split into six groups (six mice per group) when the tumor size reached approximately 100 mm³. PBS (control), regorafenib, DFMO, DR, NPs, and Man-NPs at 5 mg/kg regorafenib and 20 mg/kg DFMO every other day for about two weeks were intravenously injected into each group. Each mouse’s tumor volume and body weight were measured to assess anti-tumor efficacy and safety in vivo. The formula used to determine tumor volumes was V (mm³) = (length × width²)/2. For ethical reasons, the mice were euthanized when the biggest tumor approached 2000 cm³. Tumor tissues and major organs were collected for additional testing. The formula used to determine the tumor growth inhibition ratio (TGI %) was TGI (%) = (1−W_treat/W_control) × 100% (W_control and W_treat represent the tumor weight of the control and treatment groups, respectively). The Shanghai General Hospital Animal Ethics Committee reviewed and approved all experiments and procedures to ensure they complied with ethical regulations.

Statistical analysis

All data were analyzed using Graphpad Prism 8 and IBM statistical package for social sciences 18. Data were presented as mean ± standard deviation, and each experiment was carried out in triplicate. The student's t-test was applied to compare the differences between the two groups. One-way analysis of variance (ANOVA) was applied when more than two groups were involved. The outcomes of animal survival were compared using the log-rank (Mantel-Cox) test. Statistically significant difference was denoted as *P < 0.05, **P < 0.01, and ***P < 0.001.