Carbon ion combined photon radiotherapy induces ferroptosis via NCOA4-mediated ferritinophagy in glioblastoma

Abstract

Glioblastoma (GBM), the most prevalent and lethal primary malignancy of the central nervous system, remains refractory to conventional photon radiotherapy due to inherent limitations in dose distribution. Although carbon ion radiotherapy offers distinct advantages, including its characteristic Bragg peak deposition and superior relative biological effectiveness, its clinical application is constrained by high costs and increased toxicity. This study explores the radiobiological interactions underlying a mixed carbon ion-photon irradiation regimen, a promising strategy in advanced particle therapy. Our findings demonstrate that combined irradiation exerts synergistic cytotoxic effects in GBM models. Mechanistic analysis reveals that this combination induces clustered DNA double-strand breaks, leading to the cytoplasmic accumulation of double-stranded DNA (dsDNA) fragments. This, in turn, activates the cGAS-STING-mediated cytosolic DNA sensing pathway, which facilitates NCOA4-FTH1 axis-driven ferritinophagy and ultimately triggering iron-dependent ferroptosis. These findings offer a new mechanistic perspective on optimizing combined particle therapy regimens for GBM treatment, with significant implications for translational applications in clinical radiation oncology.

Graphical abstract

Carbon ion combined photon radiotherapy enhances cell lethality in glioblastoma by inducing ferroptosis through activation of STING and triggering the NCOA4-mediated ferritinophagy pathway.

Highlights

• •GBM Resistance to Conventional Radiotherapy: Glioblastoma (GBM) remains highly resistant to photon radiotherapy due to inherent dose distribution limitations.
• •Advantages and Challenges of Carbon Ion Therapy: Carbon ion radiotherapy offers superior biological effectiveness but is limited by high costs and increased toxicity.
• •Synergistic Effects of Sequential Irradiation: A sequential carbon ion-photon irradiation strategy enhances GBM radiosensitivity through synergistic cytotoxic effects.
• •Mechanistic Insights into Ferroptosis Induction: The combined irradiation induces clustered DNA double-strand breaks, leading to dsDNA accumulation in the cytoplasm. This activates the cGAS-STING pathway, which facilitates ferritinophagy via the NCOA4-FTH1 axis, ultimately triggering ferroptosis.
• •Clinical Implications: These findings provide a new mechanistic framework for optimizing combined particle therapy regimens, offering translational potential in clinical radiation oncology.

1. Introduction

Glioblastoma is the most aggressive and lethal primary brain tumor, characterized by rapid progression and resistance to conventional therapies [1]. Standard treatment protocols typically include surgical resection followed by photon radiotherapy and temozolomide chemotherapy. However, despite these interventions, the prognosis for glioblastoma remains poor, with a median survival time of less than 2 years after diagnosis [2,3]. This dismal outcome highlights the urgent need for more effective treatment modalities that can overcome the limitations of traditional therapies.

Carbon-ion radiotherapy (CIRT) is currently one of the most advanced radiotherapy technologies internationally. CIRT, as a form of charged particle therapy, utilizes heavy ions, which possess unparalleled advantages in the field of tumor treatment [4,5]. (a) Radiation Physics: Unlike conventional photon radiation, carbon ions are high-LET (Linear Energy Transfer) particles. Upon entering tissues, carbon ions deliver a low dose until they reach a specific depth. At this depth, the carbon ion radiation forms a sharply defined peak, known as the Bragg peak, after which the dose significantly drops [5]. This precise dose localization allows for an increase in the dose delivered to the tumor, while potentially providing better protection to surrounding healthy tissues [6]. Thus, the dose distribution pattern of carbon ion radiation concentrates the majority of the radiation dose within the tumor region, improving treatment accuracy and enhancing tumor-killing effects, while minimizing adverse reactions in surrounding normal tissues. (b) Radiation Biology: In terms of radiobiology, carbon ion radiation differs from conventional photon radiation by having a higher relative biological effectiveness (RBE). It induces more severe DNA double-strand breaks (DSBs), weakening the tumor cells' ability to repair damage [7,8]. Additionally, carbon ion exhibit a lower oxygen enhancement ratio (OER), making them particularly effective against radioresistant, insensitive, hypoxic tumor cells, and cancer stem cells, with unique tumor-killing effects [9]. Therefore, CIRT can address some of the limitations and drawbacks of conventional photon radiotherapy, offering a significant advantage, especially in tumors that are radioresistant or located in anatomically challenging regions.

Despite its advantages, high-dose carbon ion therapy is clinically impractical due to excessive costs and the risk of brain tissue necrosis [10,11]. To mitigate these challenges, a hybrid treatment approach combining carbon ion and photon radiotherapy has been proposed, aiming to optimize therapeutic efficacy while minimizing adverse effects [12], aiming to provide more precise treatment for gliomas while maximizing patient benefit.

In this study, we investigate the cell lethality of combined CIRT and photon radiotherapy on glioblastoma cells, with a particular focus on ferroptosis, a form of programmed cell death characterized by lipid peroxidation and iron metabolism dysregulation. Recent studies have indicated that ferroptosis plays a critical role in enhancing the therapeutic efficacy of cancer treatments by overcoming traditional resistance mechanisms [13]. However, the role of ferroptosis in glioblastoma treatment, especially in response to combined CIRT and photon therapy, remains poorly understood.

The goal of this study is to evaluate whether the combination of CIRT and photon radiotherapy induces enhanced cytotoxicity in glioblastoma cells through mechanisms involving ferroptosis. To this end, we performed in vitro and in vivo experiments to assess tumor cell death, DNA damage, and oxidative stress in response to combined radiotherapy. By elucidating the molecular pathways involved in the cytotoxicity of combined CIRT and photon radiotherapy, this study aims to provide a scientific rationale for the potential clinical application of this combination therapy in glioblastoma.

2. Methods

2.1. Cell culture

The glioblastoma cell line U251, U373 and A172, a widely used model in glioblastoma research, was obtained from the American Type Culture Collection (ATCC). Cells were maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10 % fetal bovine serum (FBS), 100 U/mL penicillin, and 100 μg/mL streptomycin. The cells were incubated at 37 °C in a humidified atmosphere containing 5 % CO₂. Cells were routinely tested for mycoplasma contamination and were used for experiments when they reached 70–80 % confluence.

2.2. Irradiation treatments

Photon irradiation was performed using the PXi Precision X-RAD 225 irradiator at a dose rate of 3.198 Gy/min (225 kV, SSD 40 cm, 13.3 mA). Carbon ion irradiation was conducted using the IONTRIS intensity-modulated raster scanning system with an energy of 156.34 MeV/u. The width of the spread-out Bragg peak (SOBP) is 5 mm, and the average LET value is 50 keV/μm. Additionally, a lung membrane body extension is used to adjust the Bragg peak. Precise positioning is carried out before each irradiation based on the markers on the cell molds to ensure that the Bragg peak of the carbon ion radiation is primarily localized to the layer where the cells are adhered.

2.3. Calculation of relative biological effectiveness (RBE)

The radiation dose-response curve is fitted using the linear-quadratic (LQ) model, with the following equation:Y = exp(-(α∗x+β∗x²)); Where Y represents the surviving fraction of cells, α and β are model parameters, and x denotes the radiation dose. The RBE is defined as the ratio of the photon and carbon ion doses required to achieve 10 % cell survival.

2.4. C11 BODIPY (581/591) staining

Lipid peroxidation was quantified using the C11-BODITY (581/591) probe and LpoUp was used as a positive control (Beyotime, China)., following the manufacturer's instructions. Briefly, the probe was prepared in a working solution at a 1:200 ratio with serum-free culture medium. 100 μl of the working solution was added to GBM cells for 30 min and then analyzed immediately with a flow cytometry.

2.5. Detection of Fe²⁺

Intracellular iron levels were measured using the FerroOrange probe (Dojindo Molecular Technologies, Japan). The GBM cells were incubated with 1 μmol/L working solution for 30 min and then analyzed by flow cytometry and cell immunofluorescence.

2.6. Co-immunoprecipitation(Co-IP)

To investigate the interaction between STING and NCOA4, Co-IP was performed using the Pierce Co-Immunoprecipitation Kit (Thermo Fisher Scientific, USA). Cell lysates were incubated with anti-STING antibody overnight at 4 °C, followed by the addition of Protein A/G agarose beads. After washing, the immune complexes were eluted, separated by SDS-PAGE, and analyzed by Western blotting for NCOA4.

2.7. Clonogenic assay

After digesting and resuspending the cells in optimal growth conditions, cell counting is performed. The cells are then seeded into 6-well plates and incubated for 24 h to allow for cell adhesion. Following this incubation period, the cells are subjected to irradiation. After irradiation, cells were incubated for 10–14 days to permit colony formation. Colonies were fixed with 4 % paraformaldehyde and stained with 0.5 % crystal violet for 30 min. Colonies containing more than 50 cells were counted under a microscope. Colony counting is performed using the GelCount colony counter for imaging and quantification. The surviving fraction (SF) was calculated by normalizing the number of colonies in the treated groups to the number in the control group.

2.8. Western blot analysis

Protein samples were extracted using RIPA buffer containing protease and phosphatase inhibitors. Equal amounts of protein were separated by SDS-PAGE and transferred to PVDF membranes. Membranes were blocked with 5 % non-fat milk in TBST and incubated with primary antibodies overnight at 4 °C. Membranes were washed and incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL) and quantified by densitometry using ImageJ software.

2.9. RNA sequencing (RNA-seq) and data analysis

RNA sequencing was performed to investigate the transcriptomic changes induced by the different irradiation treatments. Total RNA was extracted from glioblastoma cells 24 h post-irradiation using the RNeasy Mini Kit (Qiagen), following the manufacturer's protocol. RNA quality and quantity were assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies) and NanoDrop spectrophotometer (Thermo Fisher Scientific,USA), respectively. Libraries for RNA-seq were prepared using the TruSeq Stranded mRNA Library Prep Kit (Illumina) and sequenced on the Illumina NovaSeq 6000 platform. Raw sequencing data were processed using FastQC for quality control, and reads were aligned to the human reference genome (GRCh38) using STAR aligner. Differential gene expression analysis was performed using DESeq2, with a false discovery rate (FDR) of less than 0.05 considered significant. Supervised clustering analysis and Gene Set Variation Analysis (GSVA) are performed using R software. The gene sets for the apoptosis pathway are downloaded from the GSEA website.

2.10. Transmission electron microscopy (TEM)

After irradiating the GBM cells for 48 h, add electron microscope fixative and scrape the cells. After centrifugation, add fresh fixative and fix the cells at room temperature for 2 h. Finally, store the fixed cells at 4 °C for preservation. After overnight storage at 4 °C, wash and infiltrate the cells, then embed them in epoxy resin and cut into sections. Cell ultrastructures were examined with Hitachi TEM system and obtain images at 80 kV.

2.11. Detection of autophagy

After irradiating the GBM cells for 48 h, digest and centrifuge the cells. Prepare the Assay Buffer according to the instructions of the CYTO-ID® Autophagy Detection Kit 2.0 (Enzo Life Sciences, US). Then, dilute the Green dye in the Assay Buffer at a 1:1000 ratio to prepare the working solution. Add 100 μl of the working solution to each tube and incubate at 37 °C for 30 min and then analyzed by flow cytometry.

2.12. Detection of lysosome

After irradiating the GBM cells for 48 h, digest and centrifuge the cells. Prepare the working solution by diluting LysoTracker dye(Beyotime, China) in culture medium at a 1:15,000 ratio. Add 100 μl of the working solution to each tube and incubate at 37 °C for 30 min and then analyzed by flow cytometry.

2.13. Detection of dsDNA

To detect cytoplasmic dsDNA, cells were fixed and permeabilized as described above, then incubated with an Double-stranded DNA antibody (sigma, USA) followed by Alexa Fluor 555-conjugated secondary antibody. Cells were imaged using a confocal laser scanning microscope (Zeiss LSM 880).

2.14. Drugs and concentrations

All the inhibitors were purchased from Selleck Chemicals (USA). The following drugs were administered immediately after cell irradiation, with a treatment duration of 48 h. Ferroptosis inhibitors included Ferrostatin-1 (Ferr-1) with 5 μM and Deferoxamine (DFO) with 50 μM. Autophagy inhibitors included 3-Methyladenine (3-MA) with 5 mM and Bafilomycin A1 (BafA1) with 50 nM. The apoptosis inhibitor ZVAD-FMK (Z-VAD) with 10 μM, and the necroptosis inhibitor Necrostatin-1 (Nec-1) with 2 μM. For mouse treatment, Liproxstatin-1 (Lipro-1) was diluted in PBS and administered intraperitoneally at a dose of 10 mg/kg. The injections were administered daily after irradiation until the experimental endpoint.

2.15. Xenograft model

Four-week-old female BALB/C nude mice (Regan Biological Company) were housed in a pathogen-free environment (22–25 °C, appropriate humidity, 12-h light/dark cycle) with standard chow and water ad libitum. Intracranial tumor models were established using a stereotactic apparatus to implant luciferase-labeled U373 cells. Tumor growth was monitored via in vivo imaging on day 9 post-implantation and days 7, 14, and 21 after irradiation.

Photon irradiation was administered on day 10 post-implantation using the same parameters as the in vitro experiments. Mice were anesthetized with isoflurane, positioned with their heads exposed, and the rest of the body shielded by lead plates.

Carbon ion irradiation was delivered using a horizontal beam with full-target coverage. Following intraperitoneal pentobarbital anesthesia, mice were fixed on a PMMA board with the tumor site marked. They were positioned laterally in an irradiation mold, with a front-facing aluminum block (with a notch aligned to the head) shielding the body. Tumor localization was confirmed with laser guidance at the isocenter. Irradiation planning was conducted using the Syngo system, targeting a 20.0 mm lateral diameter and 20 mm SOBP. A PMMA range shifter was used for dose uniformity. Final carbon ion energy ranged from 118.41 to 140.01 MeV/u, with a dose-averaged LET of 96.94 keV/μm in the target region.

2.16. Immunohistochemical staining and H&E staining

First, the mouse brain tissue was extracted. After intraperitoneal anesthesia with pentobarbital, the chest cavity was opened along the sternum to fully expose the heart. After opening the left ventricle, the perfusion needle was inserted into the aorta. After securing the needle, the right atrium was cut open, and the blood was rapidly flushed with physiological saline until the outflowing liquid became clear. Subsequently, a constant infusion of 4 % paraformaldehyde was administered via an infusion pump until the liver was swollen and hardened. After the heart perfusion was complete, the skin over the head was carefully opened along the midline of the skull, and the skull was cautiously cut to remove the entire brain. The brain was then fixed in 4 % paraformaldehyde for further use. Immunohistochemical staining of primary tumors and H&E staining of tumor tissues were performed as described previously with the following antibodies overnight: 4-hydroxy-2-noneal (4-HNE, 1:200, Abcam, ab46545) and IBA-1(1:1000, Cell Signaling Technology, 17198).

2.17. Statistical analysis

All experiments were performed in triplicate unless otherwise noted. Data are presented as mean ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism software. Differences between groups were evaluated using one-way ANOVA followed by Tukey's post hoc test for multiple comparisons or by unpaired t-tests when comparing two groups. A p-value of <0.05 was considered statistically significant.

3. Results

3.1. Carbon ion combined photon radiotherapy enhances cell lethality on glioblastoma cells

The irradiation methods using photons and carbon ions are illustrated in Fig. 1A and B. First, to get the RBE for the carbon ion VS photons in GBM cell lines were calculate through colony assay. Using LQ model, the RBE₁₀ of the U251 cell line was approximately 1.6, while that of the U373 and A172 cell line was approximately 2 (Figure S1A-D and Figure S2A-B). Based on these values, 5 Gy of carbon ion irradiation was used, while photon doses of 8 Gy for U251 cells and 10 Gy for U373 and A172 cells were selected for most subsequent experiments. For the combined treatment group, each modality contributed half of its respective dose, with carbon ion irradiation followed immediately by photons. That's mean, U251 cells received 2.5 Gy of carbon ions plus 4 Gy of photons, while U373 and A172 cells received 2.5 Gy of carbon ions plus 5 Gy of photons.

Fig. 1.

Carbon ion combined with photon radiotherapy effectively inhibits the proliferation of glioblastoma cells. (A) Schematic diagram of carbon ion irradiation of cells. (B) Schematic diagram of photon irradiation of cells. (C) Colony formation results of GBM cell lines after different irradiation treatments. (D) Immunofluorescence assay detecting the expression of γ-H2AX in each group post-irradiation. (E) Flow cytometry analysis of γ-H2AX expression in each group post-irradiation. (F) Western blot analysis of phosphorylated γ-H2AX protein expression in each group post-irradiation. Scale bar: 20 μm. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Next, we investigated the impact of different irradiation modalities on GBM cell proliferation. The results demonstrated that, compared to other groups, the combined group more effectively suppressed GBM cell survival (Fig. 1C and Fig. S2C), indicating superior therapeutic efficacy and highlighting the benefit of multimodal irradiation. Further analysis revealed that combined treatment group significantly increased apoptosis (Fig. S1E) and DNA damage, evidenced by elevated γ-H2AX foci in immunofluorescence images (Fig. 1D and Fig. S2D–E) and confirmed by flow cytometry and Western blot (Fig. 1E–F and Fig. S2F–G). These findings collectively support the enhanced cytotoxicity of the combined irradiation strategy.

3.2. Carbon ion combined photon radiotherapy induces ferroptosis in glioblastoma cells

To explore the molecular basis of enhanced cytotoxicity from combined radiotherapy in glioblastoma cells, RNA-seq was performed, revealing significant enrichment of ferroptosis-related pathways in the combination group (Fig. 2A), This suggested ferroptosis as a key mechanism of cell death. To validate this, inhibitors targeting ferroptosis (ferrostatin-1), apoptosis (Z-VAD-fmk), necroptosis (necrostatin-1s), and autophagy (3-MA) were tested. And ferrostatin-1 significantly improved cell survival post-irradiation compared to other inhibitors (Figure S3A-B and Figure S4), indicating ferroptosis as the dominant mode of cell death.

Fig. 2.

Carbon ion combined with photon radiotherapy induces glioblastoma cell death by promoting ferroptosis. (A) GSVA heatmap of RNA-seq. (B) Flow cytometry to detect the level of lipid peroxidation in cells. (C) MDA assay to measure the level of lipid peroxidation in cells. (D) Flow cytometry to detect the level of Fe²⁺ in cells. (E) PI staining to assess cell death. (F) Transmission electron microscopy (TEM) analysis of changes in mitochondria and autophagosomes in each group of cells. Red arrows indicate mitochondria, and yellow arrows indicate autolysosomes. Scale bar (up): 2 μm, Scale bar (down): 500 nm. (G) Western blot analysis of ACSL4 and SLC7A11 expression in GBM cell lines. (H) Cell immunofluorescence to detect the level of unstable iron ions in cells. (I–J) Flow cytometry to detect the level of Fe²⁺ in cells. (K–N) Flow cytometry to detect the level of lipid peroxidation in cells. Ferr-1: 5 μM; DFO: 50 μM. Ferroptosis inhibitors were applied immediately after irradiation and maintained for 48 h. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

To support this, combined treatment markedly elevated lipid peroxides, Fe²⁺ levels, and cell death (Fig. 2B–E, Figure S3C-D and Figure S5A-C), indicating a stronger induction of ferroptosis. Transmission electron microscopy (TEM) also revealed typical ultrastructural features—namely,shrunken mitochondria with increased membrane density and outer membrane disruption—in the combined treatment group (Fig. 2F). Such changes were relatively infrequent in cells subjected to photon or carbon ion irradiation alone. To further assess mitochondrial damage, we measured mitochondrial membrane potential, which was significantly lower in the combined treatment group, aligning with TEM results (Fig. S6A–B). Furthermore, mitochondrial ROS, measured by MitoSOX, was increased in the combination group (Fig. S6C–D). Western blot and qPCR analyses of key ferroptosis markers further corroborated these findings. The combined treatment group showed significantly elevated ACSL4 expression and reduced SLC7A11 levels (Fig. 2G, Figure S3E-F and Figure S5D). qPCR results also revealed a significant increase in CHAC1 and PTGS2 expression in the combined treatment group (Fig. S3E-F).

To further verify that ferroptosis is a major contributor to the enhanced cell lethality observed in the combined therapy, we employed the ferroptosis inhibitors deferoxamine (DFO) and ferrostatin-1 (Ferr-1) in follow-up experiments. As demonstrated in Fig. 2H–N, both inhibitors significantly improved the level of Fe²⁺ and lipid peroxides in GBM cells exposed to the combination of CIRT and photon therapy, supporting ferroptosis as a key mechanism of cell death in this context.

To evaluate the in vivo efficacy of CIRT and photon radiotherapy, an orthotopic glioblastoma xenograft model was established in nude mice (Fig. 3A–C). Tumor progression was monitored by bioluminescence imaging (Fig. 3D–E). While XRT and CIRT alone moderately reduced tumor burden, the combination treatment resulted in the greatest suppression, as indicated by the lowest bioluminescent signal. These effects were attenuated by co-treatment with the ferroptosis inhibitor Liproxstatin-1, suggesting a ferroptosis-dependent mechanism. Histological analysis further confirmed the therapeutic benefit. H&E staining showed pronounced tumor cell damage in the combination group, which was partially reversed by ferroptosis inhibition (Fig. 3F). IHC staining revealed elevated expression of the lipid peroxidation marker 4-HNE in the combined treatment group, which was suppressed by Liproxstatin-1 (Fig. 3G). Additionally, Iba-1 expression in the hippocampus—a marker of microglial activation and neuroinflammation—was lowest in the combination group (Fig. 3G), suggesting reduced treatment-related brain injury. Together, these results support that the combined CIRT and photon radiotherapy induces ferroptosis in vivo, contributing to enhanced tumor suppression with limited neurotoxicity.

Fig. 3.

Carbon ion combined with photon radiotherapy inhibits glioblastoma tumor growth in vivo by promoting ferroptosis. (A) Schematic diagram of mouse immobilization for irradiation. (B) Illustration of carbon ion irradiation in mice. (C) Illustration of photon irradiation in mice. (D) In vivo imaging of intracranial tumors in mice (n = 7). Irradiation was performed on day 10 after orthotopic tumor implantation, followed by daily intraperitoneal injection of Lipro-1 (10 mg/kg). The control group received an equivalent volume of normal saline. (E) Quantification of in vivo imaging on day 31 after orthotopic tumor implantation. (F) HE staining of mouse brain tumors. (G) 4HNE and Iba-1 staining of mouse brain tumors. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

3.3. Carbon ion combined photon radiotherapy promotes ferroptosis by enhancing ferritinophagy

To further investigate the contribution of ferroptosis to the cell lethality of combined CIRT and XRT in glioblastoma cells, we analyzed transcriptomic data. As shown in Fig. 4A, NCOA4 expression was significantly upregulated in both U251 and U373 cells in the combined group compared to single-treatment and control groups. Given that NCOA4 is a key regulator of ferritinophagy—a selective form of autophagy that degrades ferritin to release iron and promote ferroptosis—we examined associated molecular changes. This upregulation of NCOA4 correlated with reduced expression of ferritin heavy chain 1 (FTH1) at both the mRNA and protein levels (Fig. 4B–D), suggesting enhanced ferritin degradation. TEM (Fig. 2F) further supported this by showing a higher abundance of autolysosomes in the combined group compared to single-irradiation groups. Consistently, Western blot analysis revealed elevated LC3B-II levels in the combination group, indicating increased autophagy (Fig. 4D), which was also confirmed by flow cytometry (Fig. 4E–F). Since ferritinophagy requires lysosomal function for iron release [14], we next evaluated lysosome levels by flow cytometry and found the highest levels in the combined group (Fig. 4G–H). Additionally, co-localization analysis showed that divalent iron ions were primarily enriched in lysosomes, with both lysosome and iron levels peaking in the combined group—consistent with enhanced ferritinophagy and ferroptosis (Fig. 4I).

Fig. 4.

Carbon ion combined photon radiotherapy promotes ferroptosis in glioblastoma cells by enhancing ferritinophagy. (A) Heatmap of the expression of ferroptosis-driving and inhibiting genes in each group. Three biological replicates per group. (B,C) RT-qPCR to detect the expression of NCOA4 and FTH1. (D) Western blot to detect the expression of NCOA4, FTH1, and LC3B. (E, F) Flow cytometry to detect the autophagy level in each group. (G, H) Flow cytometry to detect the lysosomal level in each group. (I) Cell immunofluorescence to detect the expression and colocalization of Fe²⁺ and lysosomes. Scar bar: 20 μm. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

To elucidate the role of autophagy in NCOA4-mediated ferritinophagy and ferroptosis, we used the autophagy inhibitors 3-methyladenine (3-MA) and bafilomycin A1 (BafA1). Both inhibitors significantly reduced lipid peroxidation in GBM cells treated with combined CIRT and photon radiotherapy, highlighting the essential role of autophagic flux in ferroptosis (Fig. S7A–D). Consistently, inhibition of autophagy also decreased intracellular Fe²⁺ levels (Fig. S7E–H), indicating that iron release from ferritin stores depends on functional autophagy. To confirm that these effects were due to disrupted ferritinophagy, we examined NCOA4 and FTH1 expression via Western blot. Both 3-MA and BafA1 treatments led to increased level of NCOA4 following irradiation, while its effect on FTH1 expression was minimal (Fig. S7I–L), likely due to impaired autophagic degradation and compensatory upregulation of cargo receptors. Previous studies have shown that NCOA4 and FTH1 form a stable complex, with FTH1 binding up to 24 NCOA4 fragments [15,16]. To validate this interaction in GBM cells, we performed endogenous Co-IP and confirmed the physical association between NCOA4 and FTH1 (Fig. S7M). These findings collectively suggest that the combined CIRT and photon therapy enhances ferroptosis in GBM cells by promoting NCOA4-dependent ferritinophagy.

3.4. Carbon ion combined photon radiotherapy promotes ferroptosis by enhancing NCOA4-mediated ferritinophagy

To further explore whether NCOA4 plays a key role in this process, we first investigate the clinical relevance of NCOA4 in glioblastoma. We analyzed patient survival data based on NCOA4 expression levels. Kaplan-Meier survival curves demonstrated that glioblastoma patients with lower NCOA4 expression had significantly shorter overall survival compared to those with higher NCOA4 expression (Fig. 5A). Additionally, high NOCA4 expression was also associated with a low grade, IDH mutation-type and 1p/19q codeletion (Fig. 5B). These findings suggest that NCOA4 may play a protective role in glioblastoma and that its expression levels are potentially associated with treatment response and patient outcomes.

Fig. 5.

Carbon ion combined photon radiotherapy promotes ferroptosis in glioblastoma cell lines by enhancing NCOA4-mediated ferritinophagy. (A) Kaplan-Meier survival curve of NCOA4 expression in glioblastoma patients. (B) Box plots of NCOA4 expression associated with various clinical pathologies of glioblastoma (WHO grade, IDH mutation status, 1p/19q co-deletion). (C) Western blot to detect the expression of NCOA4 and FTH1 in GBM cells after NCOA4 knockdown. (D) Western blot analysis of NCOA4 and FTH1 protein expression in NCOA4-knockdown cells. (E) Flow cytometry to detect the level of lipid peroxidation in NCOA4-knockdown cells. (F) Flow cytometry to detect the level of Fe²⁺ in NCOA4-knockdown cells. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Next, we explored the functional role of NCOA4 in the enhanced cell lethality of combined CIRT and photon radiotherapy by knocking down NCOA4 in glioblastoma cells using two different short hairpin RNA (shRNA) constructs (sh-NCOA4#1 and sh-NCOA4#2). Western blot analysis confirmed efficient knockdown of NCOA4 expression in GBM cells and also revealed higher levels of FTH1 in NCOA4-knockdown cells compared to control cells (sh-NC), indicating that ferritin degradation was impaired when NCOA4 expression was reduced (Fig. 5C). We then subjected these NCOA4-knockdown cells to the same radiotherapy treatments used previously. Western blot analysis revealed that the increased NCOA4 expression induced by the combined treatment was effectively suppressed upon NCOA4 knockdown. In contrast, the expression of FTH1 was significantly upregulated after NCOA4 inhibition (Fig. 5D). Additionally, we validated ferroptosis-related markers in these cells and observed that the elevated levels of lipid peroxidation and Fe²⁺ induced by the combined group were reversed (Fig. 5E–F). Based on these findings, we conclude that ferroptosis in GBM cells induced by the combination of carbon ions and photons is primarily mediated through the NCOA4-dependent ferroautophagy pathway.

3.5. Carbon ion combined photon radiotherapy promotes ferroptosis by activating STING and triggering the NCOA4-mediated ferritinophagy pathway

Based on the sequencing results, we also found that the cytosolic DNA sensor (KEGG_CYSTEINE_AND_METHIONINE_METABOLISM) pathway was significantly enriched in the combined group (Fig. 2A). Activation of the cytosolic DNA sensor cGAS is triggered by cytosolic double-stranded DNA (dsDNA) [17]., which accumulates following irradiation-induced nuclear DNA damage. Immunofluorescence staining showed markedly increased dsDNA levels in U251 and U373 cells after combined CIRT and photon radiotherapy compared to single treatments (Fig. 6A). Consistently, levels of 8-OHdG, a marker of oxidative DNA damage, were significantly elevated in the combined group (Fig. 6B), indicating increased ROS generation. This aligns with the higher lipid peroxidation previously observed. Moreover, key proteins in the cGAS-STING pathway were significantly upregulated following combined treatment (Fig. 6C), confirming activation of this signaling cascade. Together, these findings suggest that combined carbon ion and photon irradiation enhances dsDNA generation and cGAS-STING pathway activation in glioblastoma cells.

Fig. 6.

Carbon ion combined photon radiotherapy promotes ferroptosis in glioblastoma cells by activating STING. (A) Immunofluorescence detection of dsDNA expression in each group after irradiation 48h. (B) ELISA detection of 8-OHdG expression in each group after irradiation 48h. (C) Western blot of cGAS-STING pathway-related proteins in glioblastoma cells after different radiation treatments. (D) Flow cytometry to detect the level of lipid peroxidation in sting-knockdown cells. (E) Flow cytometry to detect the level of Fe²⁺ in sting-knockdown cells. (F) Flow cytometry to detect the level of lipid peroxidation in cells treated with the STING inhibitor C-176 (1 μM), with treatment immediately after irradiation and lasting for 48 h. (G). Flow cytometry to detect the level of Fe²⁺ in C-176 treated cells. (H, I) Western blot analysis of NCOA4 and FTH1 protein expression in sting-knockdown cells. (J, K) Western blot analysis of NCOA4 and FTH1 protein expression in C-176 treated cells. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

According to Jin et al., STING can induce ferroptosis through its interaction with NCOA4 [18]. To determine whether STING mediates the ferroptosis induced by combined CIRT and photon radiotherapy in GBM cells, we genetically silenced STING and found that the treatment-induced lipid peroxidation and Fe²⁺ levels were partially reduced (Fig. 6D–E). Pharmacological inhibition of STING using C-176 produced similar results (Fig. 6F–G). Western blot analysis further showed that STING inhibition suppressed NCOA4 upregulation and significantly increased FTH1 expression (Fig. 6H–K), indicating a disruption of ferritinophagy.

We next explored whether STING regulates ferroptosis through NCOA4. In STING knockdown cells, NCOA4 expression was downregulated at both the mRNA and protein levels (Fig. S8A–B). Co-immunoprecipitation confirmed a direct interaction between STING and NCOA4 (Fig. S8C). Moreover, overexpression of NCOA4 in STING-deficient cells partially restored the reduced lipid peroxidation and Fe²⁺ levels following 8Gy photon irradiation, which had been suppressed by STING knockdown (Fig. S8D–G). These results suggest that in GBM cells, radiation-induced ferroptosis is mediated by STING through regulation of NCOA4-dependent ferritinophagy.

4. Discussion

This study demonstrates that the combination of carbon ion and photon radiotherapy exerts a significantly enhanced cytotoxic effect on GBM cells compared to either treatment alone. We observed greater tumor suppression in vivo and increased DNA damage and ferroptosis in vitro, supporting the hypothesis that the combination of these therapies synergistically induces glioblastoma cell death. Notably, we found that the combined treatment promoted ferroptosis through activating STING and triggering the NCOA4-mediated ferritinophagy pathway. In vivo experiments further revealed that the combined treatment resulted in reduced tumor growth, correlating with the activation of ferroptosis-related pathways (Fig. 7).

Fig. 7.

Carbon ion combined photon radiotherapy enhances cell lethality in glioblastoma by inducing ferroptosis through activation of STING and triggering the NCOA4-mediated ferritinophagy pathway.

Radiotherapy exerts its antitumor effects primarily by inducing ionization and generating free radicals, which damage cellular components, particularly DNA [19]. Compared to photons, heavy ions such as carbon ions produce more clustered and complex DNA damage, leading to increased chromosomal aberrations and cell death. Clinical studies have confirmed that DNA damage foci in tumor tissues are significantly more abundant following carbon ion irradiation than after photon exposure [20]. In line with these findings, our results demonstrate that carbon ions induce more extensive DNA damage than photons in GBM cells. Moreover, the combined group further amplified DNA damage and significantly enhanced radiosensitivity. These results suggest that combining CIRT with photon therapy may overcome the limitations of either modality alone by maximizing irreparable DNA damage in radioresistant tumors. As for the choice of irradiation sequence (CIRT followed by XRT), this design was based on our clinical research practice, in which patients typically receive carbon ion irradiation first, followed by proton or photon therapy. Therefore, we adopted this sequence in our experiments to better mimic the clinical protocol and ensure translational relevance.

In recent years, ferroptosis has attracted increasing attention as a novel form of programmed cell death that may enhance the efficacy of traditional cancer therapies. Radiation induces not only DNA damage but also excessive ROS through water radiolysis, leading to lipid peroxidation, a key trigger of ferroptosis [21]. Several studies have demonstrated that radiation can inhibit tumor progression by inducing ferroptosis in various cancers, including lung, ovarian, and breast cancers [22,23]. Consistently, our findings reveal that the combination of carbon ion and photon therapy activates ferroptosis, providing new insights into its role in radiosensitization.

Notably, the combination treatment also induces significant apoptosis, which does not contradict ferroptosis being the predominant mode of cell death. Radiation therapy can simultaneously trigger multiple programmed cell death pathways, including ferroptosis and apoptosis, which may coexist and synergize to enhance tumor cell killing. Intracellular iron accumulation and ROS generation promote lipid peroxidation-driven ferroptosis and can also initiate mitochondrial apoptosis [24]. Key regulators such as p53 and ROS may serve as molecular links between these pathways. Therefore, we propose that the combined irradiation primarily induces ferroptosis while concurrently activating apoptosis, collectively enhancing radiotherapeutic efficacy.

An essential mechanism underlying ferroptosis is NCOA4-mediated ferritinophagy, which selectively degrades ferritin, the primary iron storage protein complex, to release bioavailable iron [[25], [26], [27]]. NCOA4 acts as a cargo receptor that delivers ferritin to autophagosomes and lysosomes for degradation; its depletion inhibits ferritin turnover and ferroptosis by limiting iron release [[28], [29], [30], [31], [32]]. Ferritinophagy plays a crucial role in maintaining cellular iron homeostasis, especially under conditions of iron demand [33,34]. Consistent with these findings, our study shows that combined carbon ion and photon radiotherapy significantly activates ferritinophagy in GBM cells, as evidenced by increased NCOA4 expression, enhanced autophagic and lysosomal activities, and elevated labile iron levels. This suggests that combination therapy promotes ferritin degradation and iron release, driving ferroptosis induction and contributing to enhanced radiosensitivity.

Radiotherapy primarily induces DNA damage, particularly double-strand breaks (DSBs), whereas ferroptosis is characterized by iron-dependent lipid peroxidation. These distinct cell death pathways raise the question of whether they are mechanistically connected. Although ferroptosis does not directly contribute to radiation-induced DNA damage, accumulating evidence suggests that DNA damage may influence ferroptosis through shared regulatory factors such as ATM and p53 [[35], [36], [37], [38]]. In our study, combined carbon ion and photon irradiation induced both extensive DNA damage and ferroptosis in GBM cells, prompting further investigation into the potential interplay between these processes.

One possible link is the release of dsDNA into the cytoplasm following radiation-induced nuclear damage. We observed a significant increase in cytoplasmic dsDNA levels after irradiation, with the combined treatment yielding the highest accumulation. Cytoplasmic dsDNA is a well-known activator of the cGAS-STING pathway, which can induce various forms of cell death, including apoptosis, necroptosis, and autophagy [17,39,40]. Recent studies also implicate STING activation in ferroptosis regulation. For example, mitochondrial DNA stress has been shown to trigger STING-dependent autophagy and ferroptosis in cancer cells [41], while STING-mediated lipid peroxidation contributes to ischemia-reperfusion injury in the intestine [18]. These findings support a mechanistic link between radiation-induced DNA damage and ferroptosis via STING signaling.

Our data further demonstrate that the ferroptosis induced by combined carbon ion-photon therapy is closely associated with STING activation in GBM cells. Interestingly, inhibition of STING alone did not abolish ferroptosis, nor did overexpression of NCOA4 alone enhance it, suggesting a synergistic requirement for both pathways. We propose that STING activation serves as a critical upstream event, while NCOA4-mediated ferritinophagy executes ferroptosis by releasing iron from ferritin stores. Together, these pathways converge to amplify ferroptotic signaling in response to radiation.

Importantly, this study identifies ferroptosis as a key mechanism underlying the enhanced cytotoxicity of combined carbon ion and photon therapy. Through upregulation of NCOA4, the treatment promotes ferritin degradation, increases labile iron levels, and drives lipid peroxidation-mediated cell death. This mechanism may offer a novel strategy to overcome the intrinsic radioresistance of glioblastoma. From a translational perspective, the combination of CIRT and photon therapy could provide therapeutic benefit in clinical settings, particularly for GBM patients with poor response to conventional photon therapy. Further preclinical and clinical studies are warranted to validate these findings and assess their therapeutic potential.

CRediT authorship contribution statement

Li Chen: Writing – original draft, Software, Resources, Formal analysis, Data curation, Conceptualization. Wanzun Lin: Writing – review & editing, Writing – original draft, Data curation, Conceptualization. Ziyu Le: Software, Project administration, Investigation. Fangzhu Wan: Validation, Supervision. Haojiong Zhang: Investigation. Shikai Geng: Visualization, Software. Qingting Huang: Formal analysis, Data curation. Jing Gao: Investigation. Wei Hu: Supervision, Software. Huaiyuan Chen: Data curation, Conceptualization. Xingyu Liu: Validation, Data curation. Chunlin Shao: Software. Fengtao Su: Supervision. Jiade J. Lu: Supervision. Lin Kong: Writing – review & editing, Funding acquisition.

Disclosure statement

All authors have declared no competing interests.

Ethics approval

The authors state that they have been approved by the Ethics Committee of Shanghai Proton and Heavy Ion Center (M-2020-002-A) and have followed the principles outlined in the Declaration of Helsinki for all animal experimental investigations.

Funding

This work was supported by the National Science and Technology Major Project (project No.2024ZD0519903, received by L.K.), the Science and Technology development Fund of Shanghai Pudong New Area (Project No.PKJ2021-Y48, received by J.G.; Project No.PKJ2024-Y60, received by Q.T.H.),the National Natural Science Foundation of China (project No.82172666 and No.82473577, received by F.T.S.), the Science Popularization Grant of Shanghai Pudong New Area (Project No.KPZ2024-02, received by Q.T.H.).

Declaration of competing interest

All authors have declared no interest interests.

Appendix A. Supplementary data

The following is the supplementary data to this article:

Data availability

No data was used for the research described in the article.