BC tissues and tumor-adjacent tissues (5 cm from the tumor margin) were obtained from 44 patients who underwent BC surgery, between 2008 and 2010, with lymph node dissection at the Chongqing Southwest Hospital (China). None of the patients underwent preoperative chemotherapy or radiation therapy. 9 from stage I breast cancer patients, 14 from stage II breast cancer patients, and 21 from stage III breast cancer patients. All study procedures were conducted with the pre-approval of the Ethics Committee of the Third Military Medical University. All the patients were informed about the sample processing steps and provided their written consent prior to the study enrollment.

1mg hollow mesoporous PEG-Fe₃O₄ (Nan Jing Jike Biotechnology, China) was mixed with 1mg C5aRA (GL Biochem, China), and the final volume of the mixture was 2mL. This mixture was kept under stirring at 4°C for 48 h. The loaded nanoparticles, PEG-Fe₃O₄@C5aRA, were decanted by centrifugation at 12,000× g for 30min.

All animal experiments were performed in compliance with institutional guidelines for use and care of animals. Wild-type (WT) BALB/C mice were obtained from the Animal Institute of Academy of Medical Science (Beijing, China). C5aR-KO mice (BALB/C background) were purchased from Jackson Laboratory (Ellsworth, Maine, US). BALB/C mice were 6-8 weeks old at the beginning of the experiments, and the groups were matched by age and sex. All mice were housed in the individual ventilated cages at the Institute of Immunology of the Army Medical University (Chongqing, China).

To evaluate the role C5a/C5aR pathway in breast cancer progression, two experimentation groups are set: 1) 4T-1 xenograft female BALB/C mice; 2) 4T-1 xenograft female C5aR-KO mice.

To figure out the effect of PEG-Fe₃O₄ on tumor progression, 4T-1 xenograft female BALB/C mice were randomly assigned into two experimentation groups: 1) treated with sterile 0.9% NaCl solution; 2) exposed to PEG-Fe₃O₄ (20 mg/kg). Mice were intravenously injected with sterile 0.9% NaCl solution or PEG-Fe₃O₄ by the lateral tail vein on day5, day7 and day9.

To test whether PEG-Fe₃O₄ could activate complement system, female BALB/C mice were randomly assigned into two experimentation groups (ten mice per group): 1) control (sterile 0.9% NaCl solution); 2) exposed to PEG-Fe₃O₄ (20 mg/kg). The selected dose of 20 mg/kg of body weight agreed with the dose applied in therapeutic studies of PEG-Fe₃O₄. We chose 90 minutes as exposure time to evidence an increment of the complement markers. Plasma concentrations of C5a was determined by ELISA with commercial kits.

Tumor model: For 4T-1 breast cancer models, 4T-1 cells (1 × 10⁶) were suspended in 0.1 ml of phosphate-buffered saline (PBS) media and injected subcutaneously into the right back of 6-8 weeks female mice on day 0.

To evaluate the potential of PEG-Fe₃O₄@C5aRA on inhibition of breast tumor growth, 4T-1 xenograft female BALB/C mice were randomly assigned into four experimentation groups: 1) control (sterile 0.9% NaCl solution); 2) exposed to PEG-Fe₃O₄ (20mg/kg); 3) exposed to C5aRA (1mg/kg); 4) exposed to PEG-Fe₃O₄@C5aRA (20 mg/kg). Tumor volumes were measured in two dimensions using a caliper and calculated by the formula (length × width²/2) (20)

For in vivo bioluminescence analysis: 4T-1-LUC cells (1 × 10⁶) were suspended in 0.1 ml of 1 × PBS media and injected into the right back of BALB/C female mice on day 0. The mice were randomly assigned into four experimentation groups: 1) control (sterile 0.9% NaCl solution); 2) exposed to PEG-Fe₃O₄ (20mg/kg); 3) exposed to C5aRA (1mg/kg); 4) exposed to PEG-Fe₃O₄@C5aRA (20 mg/kg). The tumor growth was regularly assessed imaging at day11, 13, 15. Luciferase activity was measured by the IVIS Spectrum in vivo imaging system (PerkinElmer) according to the manufacturer’s instructions.

The tumor-bearing mice were intravenously injected with PEG-Fe₃O₄@C5aRA (20 mg/kg), after injection for 4 hours, MR imaging was conducted on a 1.0-T clinical MRI scanner (Aspect M7, Israel).

The tumor-bearing mice were intravenously injected with PEG-Fe₃O₄@C5aRA (20 mg/kg), after injection for 48 hours, mice serum was obtained to test the main liver function index GOT (Sangon Biotech, Cat#D799581-0050) and GPT (Sangon Biotech, Cat#D799579-0050) according to the manufacturer’s instructions.

After different treatments, mice from different groups were euthanized by CO₂ inhalation. The main organs of the mice, such as their hearts, livers, spleens, kidneys, lungs and tumors, were collected and fixed in 10% neutral buffered formalin for 48 h. Then, these tissues were embedded in paraffin after being processed by alcohol and xylene and sectioned into 5 μm thickness.

Tumor Tissue sections were incubated with a primary antibody against NRF2 (Proteintech, Cat#16396-1-AP, Rabbit), Glutathione peroxidase 4 (GPX4) (Proteintech, Cat#67763-1-Ig, mouse), C5aR (Proteintech, Cat#21316-1-AP, Rabbit), C5b-9 (Abcam, Cat#ab55811, Rabbit), C5 (Beyotime, Cat#AF6360, Rabbit). HRP-labeled Goat Anti-Rabbit IgG(H+L) (Beyotime, Cat#A0208) and HRP-labeled Goat Anti-Mouse IgG(H+L) (Beyotime, Cat#A0216) as secondary antibody. The peroxidase activity was visualized with diaminobenzidine tetrahydroxy chloride (DAB) solution. The sections were counter stained with hematoxylin. Dark brown staining was considered positive.

Organ tissues were stained with haematoxylin and eosin (C0105, Beyotime) according to the manufacturer’s instructions.

Tumor tissue sections were stained for iron using Prussian Blue Iron Stain Kit (Solarbio, Cat#G1424) following manufacturer’s instructions. Perls’ blue staining was further enhanced using the DAB peroxidase substrate kit SK-4100 (Vector Labs).

Tumor tissues were digested with Tumor Dissociation Kit (Miltenyi Biotec, Cat#130-096-730) using gentleMACS™ Octo Dissociator with Heaters (Miltenyi Biotec) and then crushed through mesh for single cell suspension. To determine macrophage polarization, cells from tumor tissues were stained with Zombie.NIR™ Dye (BioLegend, Cat#B323327), anti-mouse CD45 (BioLegend, Cat#103132), CD11b (BioLegend, Cat#101263), F4/80 (BioLegend, Cat#123110), I-A/I-E (BioLegend, Cat#107605), and CD206 (BioLegend, Cat#141717). After staining process by following the manufacturer’s instructions, samples were analyzed by flow cytometry (CytoFLEXTM).

The human MCF-7 and MCF10A lines were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The mouse 4T-1 and HC11 cell lines were obtained from Procell Life Science and Technology. MCF-7, MCF-10A and 4T-1 were routinely cultured in Dulbecco’s Modified Eagle’s Medium (high glucose) (Gibco, Life Technologies, USA) supplemented with 10% fetal bovine serum (Gibco, Life Technologies, USA) and maintained at 37°C in a humidified incubator with 5% CO₂. HC11 was cultured in RPMI-1640 Medium (Gibco, Life Technologies, USA) supplemented with 10% fetal bovine serum (Gibco, Life Technologies, USA) and maintained at 37°C in a humidified incubator with 5% CO₂.

To investigate the role of C5a/C5aR pathway on NRF2 expression, 4T-1 was stimulated with 480 ng/mL C5a (Biovision, USA) for 30 min. For the other group, before stimulation with C5a, these cells were pre-treated with 10 nM C5aRA for 1 hour. The expression of NRF2 was assayed by western blot.

To investigate the role of C5a/C5aR pathway on ferroptosis resistance, four experiments groups are set. 1) control group; 2) the 4T-1 cells were treated with 2ug/mL Sorafenib (Targetmol, Cat# 284461-73-0) for 48 hours; 3) the 4T-1 cells were stimulated with 2ug/mL Sorafenib and 480 ng/mL C5a; 4) before stimulation with 480 ng/mL C5a and 2ug/mL Sorafenib, these cells were pre-treated with 10 nM C5aRA for 1 hour. Cell viability and intracellular ROS was measured. The expression of GPX4 and NRF2 were evaluated by western blot.

To determine the role of C5a/C5aR pathway on macrophage polarization, RAW264.7 or PMA (100ng/mL PMA, 48h) (Sigma, Cat#P1585-1MG) pre-treated THP-1 was stimulated with 480 ng/mL C5a (Biovision, USA) for 30 min. For the other group, before stimulation with C5a, these cells were pre-treated with 10 nM C5aRA for 1 hour.

To evaluate the influence of Fe₃O₄ on cell, 200ug/mL or 400ug/mL Fe₃O₄ was treated on cells for 48h, and then cell viability was assayed. The expression of GPX4 and NRF2 were evaluated by western blot.

To evaluate the potential role of PEG-Fe₃O₄@C5aRA on 4T-1, four experiments groups are set. 1) control group; 2) the 4T-1 cells were treated with 200ug/mL PEG-Fe₃O₄ for 48 hours; 3) the 4T-1 cells were stimulated with 480 ng/mL C5a (Biovision, USA) and 200ug/mL PEG-Fe₃O₄ for 48 hours; 4) the 4T-1 cells were stimulated with 200ug/mL PEG-Fe₃O₄@C5aRA, and after 1 hour, 480 ng/mL C5a was added. The intracellular ROS was measured and the expression of GPX4 and NRF2 was evaluated by western blot.

To investigate the role of PEG-Fe₃O₄ on macrophage polarization, RAW264.7 or PMA (100ng/mL PMA, 48h) pre-treated THP-1 was stimulated with 200ug/mL PEG-Fe₃O₄ for 48 hours.

To evaluate the potential role of PEG-Fe₃O₄@C5aRA on macrophage polarization, four experiments groups are set. 1) control group; 2) the cells were treated with 200ug/mL PEG-Fe₃O₄ for 48 hours; 3) the cells were stimulated with 480 ng/mL C5a and 200ug/mL PEG-Fe₃O₄ for 48 hours; 4) the 4T-1 cells were stimulated with 200ug/mL PEG-Fe₃O₄@C5aRA for 1 hour and 480 ng/mL C5a was added for 48 hours.

After different treatments, cells were replaced with fresh medium containing 1 mg/mL MTT (MCE, Cat#HY-Y0320) and then incubated at 37°C for 2 h. Crystals were dissolved in dimethyl sulfoxide, and the optical absorbance at 570 nm was measured.

To measure intracellular ROS level, cells (5 × 10⁵ per well) were seeded in 12-well plates. After different treatments, cells were detached with trypsin, washed twice with PBS, and then incubated with 10 µM DCFH-DA (Beyotime, Cat#S0033S) at 37°C for 20 min in the dark, mix upside down every 3-5 minutes. After washing with PBS for three times. Change of fluorescence intensity in the solution was measured upon excitation wavelength at 488 nm and the emission spectra was recorded from 500 to 600 nm. The fluorescence intensity at 525 nm (λmax) was plotted against the irradiation time. Rosup (500 μg/mL) acts as a positive control for ROS production.

Cells or tumor tissues were lysed in T-PER™ Tissue Protein Extraction Reagent (Thermo, Cat#78510) containing protease inhibitors cocktail (CWBIO, Cat#CW2200S). Lysates were centrifuged at 12,000 rpm for 15 min. Protein extracts were solubilized in loading buffer (Bioground, Cat#BG0022). Equal amounts of lysate were separated on SDS-PAGE (ExpressPLUS™, Cat#M42015C) and transferred onto a 0.45μm polyvinyl difluoride (PVDF) membrane (Beyotime, Cat#FFP39). The protein was identified by incubating the membrane with primary antibodies followed by horseradish peroxidase-conjugated secondary antibodies. The primary antibodies contain NRF2 (Proteintech, Cat#16396-1-AP, Rabbit, 1:1000), GPX4 (Proteintech, Cat#67763-1-Ig, mouse, 1:1000), β-actin (Beyotime, Cat#AF5003, Rabbit). HRP-labeled Goat Anti-Rabbit IgG(H+L) (Beyotime, Cat#A0208) and HRP-labeled Goat Anti-Mouse IgG(H+L) (Beyotime, Cat#A0216) as secondary antibodies.

After treatment, total RNA was extracted using RNA Extraction Kit (Beyotime, Cat#R0026). Purified RNA was converted into cDNA by reverse transcription. RT-qPCR was performed with SYBR Green Master Mix (Invitrogen) using the StepOnePlus™ Real-Time PCR system (Life Technologies, Carlsbad, CA, USA). Expression levels were normalized with internal control β-actin. The primer sequences of mouse iNOS: Forward 5’-3’CTGCAGCACTTGGATCAGGAACCTG,Reverse3’-5’GGGAGTAGCCTGTGTGCACCTGGAA; mouse Arg1: Forward 5’-3’CAGAAGAATGGAAGAGTCAG, Reverse3’-5’CAGATATGCAGGGAGTCACC; mouse mouse CCL3: Forward 5’-3’CATATGGAGCTGACACCCCG, Reverse3’-5’GAGCAAAGGCTGCTGGTTTC; human CCR7: Forward 5’-3’TGGTGGTGGCTCTCCTTGTC, Reverse3’-5’TGTGGTGTTGTCTCCGATGTAATC; human CCL17: Forward 5’-3’CGGGACTACCTGGGACCTC, Reverse3’-5’CCTCACTGTGGCTCTTCTTCG.

Statistical analysis was conducted using GraphPad Prism (version 8.0) software. The statistical significance for comparisons of two groups was evaluated using the student’s t-test. Statistical significance for comparisons of three or four groups was determined by one-way ANOVA. All data were presented as the mean ± standard error of mean (SEM). p < 0.05 was considered statistically significant. Statistical significance was denoted (no significance [NS], *p < 0.05; **p <0.01; ***p < 0.001, ****p<0.0001.) in the figures and figure legends. All experiments were independently repeated at least thrice.

To explore the role of complement in breast cancer, the C5b-9 and C5aR expression were analyzed in the tumoral tissues from BC patients at different clinical stages by IHC. C5b-9 and C5aR expression were significantly higher in the tumoral tissues than in the BC-adjacent non-tumoral tissues ( Figure 2A ). As previous study has proved that C5a/C5aR pathway could up-regulate the key ferroptosis related gene NRF2 (19), we speculated that it may regulate iron metabolism. For investigating the relationship between C5a/C5aR pathway and iron metabolism, 4T-1 cells were treatment with C5a. NRF2 was observed up-regulated expression after C5a administration, and the up-regulation was reversed after treatment with C5aRA ( Figure 2B ). Since C5a/C5aR pathway induced upregulated NRF2 expression, and previous studies have proved that NRF2 is a key regulator of the cellular antioxidant response and ferroptosis (21), we then investigated whether C5a/C5aR pathway could even partially induce ferroptosis resistance. Hence, ferroptosis on 4T-1 cells was induced by Sorafenib, a ferroptosis inducer (22). As presented in Figure 2C , 4T-1 cell death induced by Sorafenib could be reversed by C5a. Besides, ferroptosis sigh ROS was reduced after C5a treatment ( Figure 2D ). The western blot results showed C5a elevated the expression of NRF2 and GPX4 that were downregulated by Sorafenib administration ( Figure 2E ). These in vitro results indicated ferroptosis resistance induced by C5a/C5aR pathway, which could partially explain how breast cancer cells overcome ferroptosis.

In our in vivo study, female BALB/C and C5aR-KO mice were transplanted with 4T-1 cells to further evaluate the role of C5a/C5aR pathway on breast cancer progression. C5aR deficiency caused a significant reduction in the tumor growth ( Figures 2F-H ). The IHC staining of xenografts tumor tissues showed that the protein levels of NRF2 and GPX4 were significantly reduced in C5aR-KO mice ( Figures 2I-K ). In conclusion, C5a/C5aR pathway could promoted breast cancer progression and lead ferroptosis resistance in vitro and in vivo.

It was reported C5a/C5aR pathway could lead tumor-associated macrophages polarization toward immunosuppressive phenotype in human ovarian cancer (23). As macrophages play a key role in shaping the immune state within the tumor microenvironment and cancer progression (24), it makes sense to test whether C5a/C5aR pathway could accelerate breast cancer progression by influencing macrophage polarization. The ex vivo results showed C5a/C5aR pathway could promote macrophage M2-like polarization in RAW264.7 and THP-1 cells ( Figures 3A-D ). The flow cytometry analysis also demonstrated that C5aR deficiency decreased M2-like macrophages but increased that of M1-phenotype macrophages among tumor associated macrophages in vivo ( Figures 3E, F ).

Taken together, all these results demonstrated that C5a/C5aR pathway promoted breast cancer progression by leading cancer cell ferroptosis resistance and macrophage polarization towards M2 phenotype. Blocking this pathway could be a therapeutic strategy for breast cancer.

As blocking C5a/C5aR pathway may inhibit breast cancer progression by reducing ferroptosis resistance. Combined C5a receptor antagonists (C5aRA) with ferroptosis inducer might be a novel therapy strategy for breast cancer.

For iron nanoparticles have been wildly used in leading cancer cell ferroptosis, and mesoporous hollow iron nanoparticles have great potential for it also can be used as drug carrier (6, 25), we decided to test whether mesoporous hollow iron nanoparticles could be combined with C5aRA for breast cancer treatment.

First, to evaluate whether the cytotoxic effect of PEG-Fe₃O₄ was specific for cancer cells, mouse breast cancer cell line 4T-1, non-malignant mouse fibroblast cell line HC11, human breast cancer cell line MCF-7, human normal mammary epithelial cell line MCF-10A were treated with various doses of PEG-Fe₃O₄ for 48h and then subjected to MTT assay ( Figures 4A-D ). Although the cell viability was decreased by PEG-Fe₃O₄ treatment both in cancer cell lines and non-malignant cell lines, PEG-Fe₃O₄ was far more toxic to cancer cells than normal cells. When treated these cells with PEG-Fe₃O₄ at the concentration of 200ug/mL, the cell viability of 4T-1 decreased 26%, but only 4% for HC11; Consistent with this result, the cell viability of MCF-7 decreased 11% but only 4% for MCF-10A. To explore whether the inhibition of cell viability was through ferroptosis, the intracellular reactive oxygen species (ROS) level in cancer cells after PEG-Fe₃O₄ treatment was determined. PEG-Fe₃O₄ treatment drastically increased the intracellular ROS level in 4T-1 cells ( Figure 4E ). Besides, NRF2 and GPX4 were both reduced in 4T-1 cells after PEG-Fe₃O₄ treatment ( Figure 4F ). Together, these findings revealed that mesoporous hollow PEG-Fe₃O₄ showed cancer cell specific cytotoxicity, and it could induce ROS generation and cancer cell ferroptosis.

Then, the in vivo anti-tumor effect of mesoporous hollow PEG-Fe₃O₄ was tested in 4T-1 subcutaneous xenograft animal models. As showed in Figures 4G-I , tumor volume and tumor weight of 4T-1 xenografts were significantly reduced after intravenous (i.v.) injection of PEG-Fe₃O₄ as compared to the 0.9% NaCl solution control. The IHC staining of xenografts tumor tissues showed that the protein levels of NRF2 and GPX4 of xenograft tumor tissues were significantly reduced by PEG-Fe₃O₄ exposure ( Figures 4J-L ).

Besides ferroptosis inducing ability, previous study also showed that mesoporous hollow PEG-Fe₃O₄ could regulate macrophages polarization in neural system (7). We evaluated the ability of PEG-Fe₃O₄ to modulate macrophage polarization in vitro. PEG-Fe₃O₄ iron oxide nanoparticles were added to RAW264.7 cells for 48h and the RT-qPCR results revealed that PEG-Fe₃O₄ treatment promoted the M1-polarization induction-derived overexpression of iNOS ( Figure 4P ), while attenuated the expression of the M2 polarization gene Arg1 ( Figure 4Q ). To further evaluate whether macrophages can be reprogrammed by PEG-Fe₃O₄, PMA pre-treated THP-1 macrophages were treated with PEG-Fe₃O₄, the RT-qPCR results also demonstrated that PEG-Fe₃O₄ was able to enhance the expression of M1 marker CCR7 but reduced the level of M2 marker CCL17 in THP-1 macrophages ( Figures 4R, S ). All these data suggested that PEG-Fe₃O₄ could modulate macrophage polarization towards M1 phenotype.

Next, to verify whether it could reprogram macrophages in breast cancer model, we collected the endpoint 4T-1 xenograft from the mice to analyze tumor-infiltrating macrophages by flow cytometry analysis ( Figure 4T ). The flow cytometry analysis showed that PEG-Fe₃O₄ treatment increased the ratio of M1-like macrophages/M2-like macrophages ( Figure 4U ).

These results demonstrated that PEG-Fe₃O₄ could inhibit breast cancer development through leading cancer cells ferroptosis and macrophage polarization to M1 phenotype. And for its drug carrier ability, loading it with C5aRA could be an effective breast cancer treatment strategy.

Despite the anti-tumor effect of PEG-Fe₃O₄, it has been reported that iron nanoparticles could induce complement system activation, which may bring negative effect on tumor inhibition (14). To test whether the mesoporous hollow PEG-Fe₃O₄ could activate C5a/C5aR pathway, female BALB/C mice were administered with PEG-Fe₃O₄. After exposure, PEG-Fe₃O₄ induced 2.9-fold increment of C5a ( Figure 4M ). Under pathological conditions, after 4T-1 xenografts mice were treated with PEG-Fe₃O₄, the IHC staining of xenografts tumor tissues showed that the level of C5b-9 of xenograft tumor tissues was significantly increased ( Figures 4N, O ). All these results showed that PEG-Fe₃O₄ could induce complement system activation under both normal physiological and pathological conditions.

Based on the forward results, we postulated that anti-C5a/C5aR pathway treatment may enhance the anti-tumor ability of PEG-Fe₃O₄. As an efficient drug delivery carrier, it is possible to load mesoporous hollow PEG-Fe₃O₄ with C5aRA. As showed in Figure 5A , Fe₃O₄@C5aRA was constructed by loading mesoporous hollow PEG-Fe₃O₄ with C5aRA. The drug loading capacity (DLC) and encapsulation efficiency (EE) of PEG-Fe₃O₄@C5aRA were quantified by reading the absorbance of C5aRA at 360 nm. The DLC and EE were calculated to be 12.29% and 70.05%. Then, the in vitro release behavior of C5aRA from PEG-Fe₃O₄@C5aRA was determined at 37°C under the microenvironment PH condition outside the tumor (PH=6.8). According to the detection data ( Figure 5E ), the release of polypeptides was concentrated in the first 8 hours. The release amount reached nearly 70% in the first 8 hours and the packaged polypeptides were slowly released over time. As displayed in Figure 5B , loading PEG-Fe₃O₄ with C5aRA had no influence on the PEG-Fe₃O₄ iron oxide nanoparticles size. The transmission electron microscopy (TEM) images showed PEG-Fe₃O₄ and PEG-Fe₃O₄@C5aRA nanoparticles with a shape close to spherical and average diameter of 220nm. Dynamic light scattering (DLS) data indicated that PEG-Fe₃O₄@C5aRA had a larger hydrodynamic diameter (283.3 nm), compared to PEG-Fe₃O₄ (244.9 nm) ( Figure 5C ). The polydispersity index (PDI) values of PEG-Fe₃O₄ (0.207) and PEG-Fe₃O₄@C5aRA (0.245) were all < 0.3, indicated that the materials had good dispersibility. The average zeta potential of PEG-Fe₃O₄@C5aRA (−10.7 mV) was lower than that of bare PEG-Fe₃O₄ (−7.45 mV), reflecting the successful loading of C5aRA on PEG-Fe₃O₄ nanoparticles ( Figure 5D ).

To test whether PEG-Fe₃O₄@C5aRA could accumulate in tumor, 4T-1 xenograft mice were intravenous injection with PEG-Fe₃O₄@C5aRA, the control group were treated with 0.9% NaCl solution. As presented in Figure 5F , an obvious darkening effect in the tumor and kidney was observed in T₂-weighted MR images. The darkening of the MR images in the mouse organs after injection was also confirmed quantitatively. The signal intensities of tumor and kidney were reduced remarkably after the administration of PEG-Fe₃O₄@C5aRA. It is no doubt PEG-Fe₃O₄@C5aRA could reach kidney, and the same change of signal intensities in tumor and kidney could demonstrate that PEG-Fe₃O₄@C5aRA accumulated in tumor 4 hours post injection ( Figures 5G, H ).

We also investigated the systematic toxicity of PEG-Fe₃O₄@C5aRA in vivo by H&E staining sections. As shown in Figure 5I , there was no observable pathological damage in the H&E staining sections in major organs of mice after different treatments. Furthermore, biochemical analysis of blood samples was conducted after different treatments for 48 h. The results revealed that the primary liver function markers: glutamic oxaloacetic transaminase (GOT) and glutamic pyruvic transaminase (GPT) ( Figure 5J ) remaining in a standard range. Those results revealed excellent biocompatibility and biosafety of PEG-Fe₃O₄@C5aRA.

To determine whether PEG-Fe₃O₄@C5aRA displayed enhanced anti-tumor effect, 4T-1 cells were treated with PEG-Fe₃O₄, PEG-Fe₃O₄ plus C5a, and PEG-Fe₃O₄@C5aRA plus C5a. The MTT result showed that PEG-Fe₃O₄@C5aRA displayed stronger tumor inhibition effect ( Figure 6A ). C5a decreased the generation of ROS induced by Fe₃O₄ treatment, and C5aRA loaded in the PEG-Fe₃O₄ could reverse it ( Figure 6B ). Besides, NRF2 and GPX4 were both reduced in 4T-1 cells after PEG-Fe₃O₄@C5aRA treatment ( Figure 6C ). Together, all these results revealed that PEG-Fe₃O₄@C5aRA which combined PEG-Fe₃O₄ and C5aRA could contribute to 4T-1 cell ferroptosis, increasing ROS generation and down-regulation of NRF2 and GPX4. Besides, treated RAW264.7 and THP-1 cells with PEG-Fe₃O₄@C5aRA could promote macrophages polarization to M1 phenotype but decreased that of M2-like macrophages ( Figures 6D-G ).

To better evaluate the potential of combined therapy in a model that more closely resembles the clinical setting, we treated established 4T-1 tumors during a defined period ( Figure 7A ). In the subcutaneous 4T-1 model, treatment with PEG-Fe₃O₄@C5aRA resulted in a significant reduction of tumor growth, as compared with PEG-Fe₃O₄ or C5aRA treatment alone. By day 15, tumor volumes in mice treated with the PEG-Fe₃O₄@C5aRA were smaller than those in control group, mice treated with PEG-Fe₃O₄ or C5aRA groups ( Figures 7B-D ). The super anti-tumor effect of PEG-Fe₃O₄@C5aRA was further verified by in vivo monitor of the tumor growth using bioluminescence imaging. 4T-1-luciferase (4T-1-LUC) cells were inoculated into the back of BALB/C mice. Tumor growth at various stages was recorded by IVIS Spectrum in vivo optical imaging system (PerkinElmer) on day 11, 13 and 15 after 4T-1-LUC cells inoculation. The results have shown that the bioluminescence signals decreased significantly in the PEG-Fe₃O₄@C5aRA treatment group compared to the control and treatment alone groups ( Figure 7E ).

The western blot and IHC staining of xenografts tumor tissues showed that the protein levels of NRF2 and GPX4 of xenograft tumor tissues were significantly reduced by PEG-Fe₃O₄@C5aRA exposure. The C5b-9 stain result also proved that PEG-Fe₃O₄ activated complement system, loading it with C5aRA inhibited the activation. ( Figures 7F-I ). And the PEG-Fe₃O₄@C5aRA treatment resulted in a significant elevation of the ratio of M1 macrophage/M2 macrophage in TME ( Figures 7J-L ). From all these experiments, we concluded that that the PEG-Fe₃O₄@C5aRA exhibited superior antitumor efficacy.