Nintedanib enhances tumor cell radiosensitivity by promoting ferroptosis and modulating the ATF4/SLC7A11/GSH axis

Cell lines

LLC, MC38, A549, and HCT116 cell lines were sourced from the Oncology Laboratory of Tongji Hospital (Huazhong University of Science and Technology, Wuhan, China). MC38 (murine colon cancer cells) and A549 (human lung cancer cells) were cultured in RPMI-1640 (Hyclone, Logan, UT, USA) and supplemented with 10% fetal bovine serum (160000-044; Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, USA) and 100 U/mL penicillin/streptomycin (15140122; Gibco, Thermo Fisher Scientific, Inc.) and maintained at 37°C in 5% CO₂. LLC (murine lung cancer cells) and HCT116 (human colon cancer cells) were cultured in Dulbecco’s modified Eagle’s medium (Hyclone) with 10% fetal bovine serum and 100 U/mL penicillin/streptomycin under the same conditions.

Reagents

Nintedanib was purchased from Selleck Chemicals (S1010; Houston, TX, USA). InVivoMab anti-mouse PD-L1 antibody was obtained from BioXcell (clone 10F.9G2, catalog #BE0101; West Lebanon, NH, USA). Z-VAD-FMK (S7023), ferrostatin-1 (S7243), necrostatin-1 (S8037) were purchased from Selleck Chemicals. N-acetyl cysteine (HY-B0215; Monmouth Junction, NJ, USA) was purchased from MCE.

Experimental animals

A total of 84 C57BL/6 mice and 84 immunocompromised (nude) mice were provided by Jiangsu GemPharmatech Co., Ltd. (Nanjing, China). Male mice were used to establish the LLC tumor-bearing mouse model and female mice were used to establish the MC38 tumor-bearing mouse model^35,36. MC38 and LLC cells (1 × 10⁶) were subcutaneously implanted into syngeneic C57BL/6 and nude mice. MC38 cells transfected with an empty vector or ATF4 plasmids were similarly implanted into nude mice. Mice were randomly assigned to four treatment groups [PBS, nintedanib (50 mg/kg, 5 days on, 1 day off, administered 10 times), RT (8 Gy), anti-PD-L1 antibody (10 mg/kg, once every 3 days, administered 3 times)] when tumor volumes reached 70–100 mm³. The tumor volume was measured using a digital caliper and calculated as V = [length × width²] × 0.5, where length represents the maximum diameter and width is the shortest diameter. Tumor growth and survival curves were generated from independent experiments. The mice were euthanized by CO₂ inhalation as a humane endpoint when the tumor volume reached 2,000 mm³. Death was confirmed by cessation of both respiratory and cardiac activity. Animal experiments were performed under standard laboratory conditions at the Tongji Hospital Laboratory Animal Center with approval from the Ethics Committee of Huazhong University of Science and Technology (Approval No. TJH-202206052).

RT

In vitro and in vivo experiments were performed using an X-ray irradiator (RS 2000 Biological System irradiator; Rad Source Technologies Inc., Boca Raton, FL, USA). Tumor-bearing mice were anesthetized by intraperitoneal (i.p.) injection of 1.5% pentobarbital sodium at a dose of 50 mg/kg body weight and received a single X-ray dose fraction (8 Gy) to the area of the subcutaneous tumor using the RS 2000 Biological System irradiator. A heating pad was used during anesthesia until the mice were fully awake. Tumor cells were irradiated at a dose ranging from 2–8 Gy.

In vivo optical imaging analysis

Mice were injected with D-fluorescein i.p. (115144-35-9; GoldBio, St. Louis, Missouri, USA) and anesthetized by i.p. injection of 1.5% pentobarbital sodium at a dose of 50 mg/kg body weight. A heating pad was used during anesthesia until the mice were fully awake. Bioluminescence was detected using an IVIS 200 Xenogen system (IVIS Spectrum; Perkin Elmer, Waltham, MA, USA).

Histologic analysis and immunohistochemistry (IHC)

Isolated tumors were fixed in 4% paraformaldehyde at 25°C for 48 h, dehydrated, embedded in paraffin wax, and cut into slices for further experiments. The slices were dewaxed, rehydrated, and underwent antigen repair. Antigen repair was performed using Tris-EDTA buffer [pH 9.0] (Servicebio, Wuhan, China) in a microwave oven at 95°C for 20 min. Next, the slices were blocked with 5% bovine serum albumin (ST023; Beyotime, Shanghai, China) for 30 min at room temperature, then incubated with primary antibodies overnight at 4°C. Primary antibodies used in the experiments included anti-CD3 (1:200, ab16669; Abcam, Cambridge, UK), anti-CD8α (1:2,000, ab217344; Abcam), and anti-Ki-67 (1:200, ab16667; Abcam). After incubation with secondary antibody (1:5,000; Aspen, Wuhan, Hubei, China), the slices were stained using 3,3′-diaminobenzidine (Aspen) and counter-stained using hematoxylin (Aspen). All slices were viewed and imaged under a microscope (Leica, Wetzlar, Germany) and the positive cells were quantified by two blinded pathologists.

TUNEL assay

Isolated tumors were fixed in 4% paraformaldehyde at 25°C for 48 h, dehydrated, embedded in paraffin wax, and cut into slices for further experiments. The slices were stained using a TUNEL assay kit (11684795910; Roche, Mannheim, Germany) following the manufacturer’s protocol. The images were captured using a microscope (Pannoramic MIDI; 3D Histech, Budapest, Hungary) and analyzed using ImageJ software.

Multiplex immunofluorescence (mIF)

Isolated tumors were fixed in 4% paraformaldehyde at 25°C for 48 h, dehydrated, embedded in paraffin wax, and cut into slices for further experiments. The slices were dewaxed, rehydrated, and underwent antigen repair. Antigen repair was performed using Tris-EDTA buffer [pH 9.0] (Servicebio) in a microwave oven at 95°C for 20 min. The slices were blocked with 5% bovine serum albumin (ST023; Beyotime) for 30 min at room temperature after permeabilization with 0.3% Triton X-100 in PBS for 10 min, then incubated with the following primary antibodies: anti-CD4 (1:2,000, ab183685; Abcam); anti-CD8α (1:2,000, ab217344; Abcam); CD31 (1:500, ab182981; Abcam); and α-SMA (1:100, AF1032; Affinity, Melbourne, AUS). After incubation with fluorescent secondary antibody and DAPI (Servicebio), the immunofluorescence images were captured using a Pannoramic MIDI and analyzed using ImageJ software.

Apoptosis analysis

Cultured cells (1 × 10⁵ cells per sample) were harvested after different treatments. An FITC Annexin V Apoptosis Detection Kit I (556547; BD Pharmingen, San Diego, California, USA) was used to detect cell apoptotic status. Flow cytometric analysis was performed using CytoFLEX LX (Beckman, Brea, CA, USA). The data were analyzed using FlowJo software (v10; BD Biosciences, San Diego, California, USA).

Colony-forming assay

Cells were seeded at a low density (200–800 cells/well) in 6-well plates and the treatment (drug or radiation) was applied the following day. After 2 weeks of culture the colonies were fixed with 4% buffered paraformaldehyde and stained with 0.01% crystal violet. Colonies containing ≥50 cells were counted and used for statistical analysis.

Cell counting Kit-8 (CCK-8) assay

Cells from different cell lines were seeded into 96-well plates (4–7 × 10³ cells per well) and incubated at 37°C in a CO₂ incubator for 24 h. Following treatment, the cells were cultured for specific time periods. Cell viability was assessed using the Cell Counting Kit-8 (HY-K0301; MedChem Express, New Jersey, USA) with absorbance measured at 450 nm using a microplate reader (BioTek, Winooski, VT, USA).

RNA-seq assay

Total RNA from MC38 cell lines was extracted using Trizol (9766; Takara Bio, Shiga, Japan) and purified for library preparation and sequencing on the Illumina Hiseq platform (San Diego, California, USA). RNA sequencing was performed by Novogene (Beijing, China). Analysis of differential gene expression among the four groups was performed with the DESeq2 R package (version 1.20.0). Statistical significance for differential expression was defined as a P < 0.05. The R ‘clusterProfiler’ package was utilized to conduct Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis on the differentially expressed genes (DEGs). The corresponding datasets have been submitted to a public database (Gene Expression Omnibus under the GSE297011 dataset; www.ncbi.nlm.nih.gov/geo).

Reactive oxygen species (ROS) detection

An ROS Assay Kit (S0034S; Beyotime) was used to detect the generation of intracellular ROS. Cells (2 × 10⁵) were seeded in 6-well plates and treated as specified. After co-incubation with 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) for 30 min at 37°C, cells were thrice-washed with cold PBS. The immunofluorescence images were observed using a Pannoramic MIDI and fluorescence intensity was measured using ImageJ.

Lipid peroxidation measurement

Cells (2 × 10⁵) were seeded in 6-well plates and treated as specified. Cultured cells (3 × 10⁵ cells per sample) were harvested and stained for 30 min at 37°C with 5 μM BODIPY 581/591 C11 dye (D3861; Invitrogen, Thermo Fisher Scientific, Inc., Waltham, MA, USA) to assess lipid peroxidation. After washing three times with cold PBS, flow cytometric analysis was performed using the CytoFLEX LX. Data were analyzed with FlowJo software (v10).

Malondialdehyde (MDA) measurement

The MDA content in tumor cells was measured using the corresponding kits (A003-1-2; Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer’s instructions.

Quantitative real-time PCR (qRT-PCR)

Cells (5 × 10⁵) were seeded on 6-well plates and treated as specified. Total RNA was extracted using Trizol and reverse-transcribed to cDNA using the PrimeScript 1st Strand cDNA Synthesis Kit (6110A; Takara Bio). RT-qPCR was performed using ChamQ Universal SYBR qPCR Master Mix (Q711-02; Vanzyme, Nanjing, China) on the ABI-7900HT Sequence Quantification System (Applied Biosystems, Foster, CA, USA), according to the manufacturer’s protocol. mRNA expression of the target genes were normalized to β-actin expression and analyzed using the 2^−ΔΔCq method. The primer sequences are listed in Table S1.

Transmission electron microscope

Cells were collected after treatment and fixed in 2.5% glutaraldehyde overnight followed by fixation in 1% osmium tetroxide. The cells were stained with 2% uranyl acetate, dehydrated, embedded, and sectioned. Images of the cells were captured using a transmission electron microscope [TEM] (HT7700; Hitachi, Tokyo, Japan).

Western blot

Total protein was extracted using RIPA buffer (Beyotime) containing 1% phenylmethylsulfonyl fluoride and 1% phosphatase inhibitor (Servicebio). The protein concentration was determined using the Bradford method. Thirty micrograms of protein sample were separated using 8%–10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to 0.45 μm polyvinylidene fluoride membrane (Millipore, Billerica, MA, USA, USA). The membrane was probed with primary antibodies overnight at 4°C after incubation with 5% bovine serum albumin for 1 h at room temperature. The membranes were visualized by SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific) after the membranes were incubated with secondary antibodies (Promoter, Wuhan, China). The signal from the blots was detected using the G:BOX Chemi X system (Syngene, Cambridge, UK) and analyzed using ImageJ software. Western blotting was performed using the following antibodies: GAPDH (1:20,000, 60004-1-lg; Proteintech, Wuhan, China); β-actin (1:5,000, 60008-1-lg; Proteintech); SLC7A11/xCT (1:1,000, A2413; ABclonal, Wuhan, China); γ-H2AX (1:1,000, #9718; Cell Signaling Technology, Danvers, Massachusetts, USA); ATF4 (1:1,000, A18687; ABclonal); ATF3 (1:1,000, #18665; Cell Signaling Technology); NRF2 (1:1,000, #12721, Cell Signaling Technology); P53 (1:1,000, #2524; Cell Signaling Technology).

Reduced glutathione (GSH) measurement

GSH content in tumor cells was quantified using the corresponding kits (A005-1-2; Nanjing Jiancheng Bioengineering Institute, China) following the manufacturer’s protocol.

Cell transfection

Lentivirus particles targeting human ATF4 (shATF4) were purchased from GeneChem (Shanghai, China), with target sequences as follows: sh1 (-CCACTCCAGATCATTCCTTTA-); sh2 (-GTTGGTCAGTCCCTCCAACAA-); and sh3 (-GCCTAGGTCTCTTAGATGATT-). Viral particles and polybrene (6 μg/mL; GeneChem) were mixed with the cell culture medium and incubated for 6 h, after which fresh medium was added. Puromycin (2 μg/mL; GeneChem) was used to screen cells with successful transfection. Plasmids containing SLC7A11 or ATF4 cDNA were utilized to generate stable SLC7A11 overexpressing human cells, ATF4 overexpressing human cells, and ATF4 overexpressing mouse cells. Transfection was performed using Lipofectamine 3000 (L3000015; Invitrogen, Waltham, MA, USA). pCDH-pCDH-CMV-MCS-ATF4 lentivirus particles were produced by co-transfecting pCDH-pCDH-CMV-MCS-ATF4 with packaging plasmids (pMD2 and pAX2) into HEK293T cells. MC38 cells were infected with pCDH-pCDH-CMV-MCS-ATF4 lentivirus particles and selected with puromycin (2 μg/mL; GeneChem) to establish a stable ATF4-overexpressing MC38 cell line. Western blot analysis confirmed the efficiency of gene knockdown or overexpression.

Flow cytometry

Cultured cells (1 × 10⁵ cells per sample) were harvested and stained with antibodies against PD-L1 (PE anti-mouse CD274, 1:100, clone 10F.9G2, catalog 124307; BioLgend, San Diego, California, USA) at 4°C for 30 min. Tumor tissues were minced and digested with type IV collagenase, hyaluronidase, and DNase (Biosharp, Anhui, China) at 37°C for 1 h, then filtered through 100-μm nylon cell strainers (Biosharp). Lymphocytes were collected, and live/dead staining was performed using eBioscience Fixable Viability Dye eFluor™ 780 (65-0865-14, 1:1,000; Invitrogen). The cell surface was stained with antibodies against CD45 (BV480 anti-mouse CD45, 1:100, clone 30-F11, catalog 566095; BD Horizon, San Diego, California, USA), CD3 (PE anti-mouse CD3, 1:100, clone 17A2, catalog 100205; BioLgend), and CD8a (APC anti-mouse CD8a, 1:100, clone 53-6.7, catalog 100711; BioLgend). The eBioscience FOXP3/Transcription Factor Staining Buffer Set (00-5523-00; Invitrogen) was used before intracellular staining with antibody against Ki-67 (PE/cyanine anti-mouse KI-67, 1:100, clone 16A8, catalog 652425; BioLgend). Flow cytometry was performed using CytoFLEX LX and the data were analyzed using FlowJo software (v10).

Safety evaluation

The main organs (heart, liver, lungs, colon, spleen, and kidneys) with serum were collected for further analysis at the end of the animal experiments. Organs were sectioned and stained with H&E for histopathologic evaluation of tissue damage. Mouse serum was collected for blood biochemistry analysis, including measurements of alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and creatinine (CRE), and analyzed using a Chemray 240 Automatic Biochemical Analyzer (Rayto Science, Shenzhen, China).

Statistical analysis

Statistical analyses were performed using GraphPad Prism 8.0 (San Diego, CA, USA). Data are presented as the mean ± SD unless specified. Repeated-measures ANOVA with Tukey’s multiple comparison test were used for animal experiments. Mouse survival curves were calculated using the Kaplan–Meier method and compared by the log-rank test. One- or two-way ANOVA was used for comparisons, as appropriate. A P value < 0.05 was considered statistically significant, with significance noted as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Nintedanib combined with RT significantly inhibits tumor growth in vivo

LLC and MC38 cell lines were used to establish tumor models in C57BL/6 mice to evaluate the antitumor efficacy of nintedanib in combination with RT (Figure 1A). The in vivo antitumor efficacy and safety of nintedanib was assessed in preliminary experiments in MC38 tumor-bearing mice using dose gradients of 30, 50, and 70 mg/kg. The 50 mg/kg dose gradient provided antitumor effects comparable to the antitumor effects achieved with 70 mg/kg. However, the 70 mg/kg regimen resulted in severe diarrhea and weight loss in tumor-bearing mice. Given the robust antitumor effects demonstrated with 2-week 50 mg/kg nintedanib administration in studies by Kutluk Cenik et al., Awasthi et al., and our own research^26,30,37 and to minimize the inevitable esophageal damage caused by gavage administration in mice, the administration frequency of nintedanib was determined to be 50 mg/kg, 5 days on, 1 day off, administered 10 times.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Antitumor effect of nintedanib combined with radiotherapy in vivo. (A) Treatment schedule of tumor-bearing mice. LLC and MC38 cells were implanted subcutaneously into C57BL/6 mice (n = 7 per group). (B, C) Tumor growth curves of tumor-bearing mice. (D, E) Survival curves of tumor-bearing mice. (F, G) In vivo fluorescence imaging of MC38 tumor-bearing mice on days 15 and 30 after different treatments. (H) Quantitative analysis of in vivo fluorescence imaging of MC38 tumor-bearing mice on day 15 after different treatments. LLC and MC38 cells were implanted subcutaneously into immunocompromised (nude) mice (n = 7 per group). (I–L) Tumor growth and survival curves of LLC and MC38 immunocompromised (nude) tumor-bearing mice. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

The tumor growth curves of C57BL/6 tumor-bearing mice indicated that the strategy of combining nintedanib with RT had significantly superior anti-tumor efficacy compared to monotherapy (Figures 1B, C and S1A, B). Moreover, the combination groups showed the best overall survival compared to other treatment groups and the control group (Figure 1D, E). In vivo optical imaging of MC38 tumors on days 15 and 30 post-treatment showed that all groups, except the combination group, had reached the endpoint (Figure 1F, G). Fluorescence imaging on day 15 and quantitative analysis further confirmed the antitumor efficacy of nintedanib combined with RT (Figure 1H). In addition, immunohistochemical staining for CD3 and CD8 revealed the highest infiltration of immune cells in the tumor periphery and center in the combination group (Figure S2A–F). Multiplex immunofluorescence staining also demonstrated an increased ratio of CD8⁺ T cells-to-CD4⁺ T cells within the tumor tissues (Figure S2G, H). Ki-67 and TUNEL staining were performed to assess cell proliferation and apoptosis in tumor tissues. As expected, Ki-67, a marker of cell proliferation, was reduced in the treatment groups, with the lowest expression observed in the combination group (Figure S2I, J). TUNEL, an indicator of apoptosis, also showed the most prominent efficacy in the combination group (Figure S2K, L). These results indicated that nintedanib combined with RT significantly increased the infiltration of immune cells in tumor tissues and had excellent antitumor effect.

Studies have demonstrated that nintedanib increase T cell and DC infiltration, inhibit cancer-associated fibroblast, and enhance immunotherapy efficacy in mice^26,27,38. LLC and MC38 tumor models were established in athymic nude mice to determine whether the antitumor effects of nintedanib depend on an intact immune system. Tumor growth in nude mice were significantly reduced in the three treatment groups with the combination treatment showing the most effective antitumor response as shown in Figure 1I–1L. The strategy of combining nintedanib with RT exerted the optimal antitumor effect in the C57BL/6 and nude mouse tumor-bearing models and the antitumor effect was more pronounced in C57BL/6 tumor-bearing mice with an intact immune system by comparing the survival (Figure 1D,1E, 1K, and 1L) and individual tumor growth curves (Figure S1) of different tumor-bearing mouse models. Furthermore, the changes in tumor growth between intact and nude tumor-bearing mice were compared at different time points. Immunocompetent tumor-bearing mice model responded more significantly to combination therapy than nude mice (Figure S2M), suggesting that the adaptive antitumor immunity is one of the important mechanisms underlying the combination of nintedanib and RT. Although an intact immune system extended tumor-bearing mice survival, the antitumor advantage of the combination treatment was still observed in immunocompromised mice, suggesting that combination treatment remains effective in the absence of an adaptive immune response. These results suggest that nintedanib has the potential to directly act on tumor cells in addition to improving the tumor immune microenvironment, thereby enhancing the anti-tumor efficacy of RT.

Nintedanib combined with RT significantly suppresses cell proliferation in vitro

To investigate the combined therapeutic potential of nintedanib, a series of in vitro experiments were performed to evaluate antitumor effects in combination with RT. The results from colony formation and CCK8 assays showed that the combination of nintedanib and radiation significantly inhibited tumor cell proliferation (Figure S3A, B). An unbiased RNA-seq analysis was performed to assess changes in gene expression induced by nintedanib and to elucidate the mechanisms underlying the antitumor effects following combination treatment. A heat map of gene expression revealed that nintedanib significantly enhanced the transcriptional changes induced by radiation (Figure S3C). Tumor-related KEGG enrichment pathways between the combined and radiation groups and between the nintedanib and control groups were compared to determine the effect of nintedanib on the proliferation and death of tumor cells, which showed significant differences in the expression of genes related to ferroptosis (Figure S3D, E). Ferroptosis is an iron-dependent, non-apoptotic form of regulated cell death caused by lipid peroxidation, which is controlled by integrated oxidation and antioxidant systems. Several metabolic pathways, particularly those involving iron, lipids, and amino acids, orchestrate the complex ferroptotic response through direct or indirect regulation of iron accumulation or lipid peroxidation³⁹. Gene Set Enrichment Analysis (GSEA) showed that nintedanib impacted the expression of genes related to lipid metabolism, GSH, and amino acid transport in tumor cells (Figure S3F–H), indicating that nintedanib may modulate ferroptosis sensitivity and affect key metabolic pathways.

Nintedanib enhances radio-sensitivity by inducing ferroptosis in tumor cells

The effects of the ferroptosis inhibitor, ferrostatin-1 (Ferr-1), the apoptosis inhibitor, Z-VAD-FMK (Z-VAD), and the necroptosis inhibitor, necrostatin-1s (Nec-1s), were evaluated in tumor cells treated with nintedanib to determine whether nintedanib induces ferroptosis in tumor cells. Various cell death inhibitors could reverse the proportion of late-stage apoptosis (Annexin V+/PI+) induced by nintedanib with ferroptosis inhibitors exhibiting the most significant reversal effect (Figure 2A). CCK8 and colony formation assays revealed that ferrostatin-1 more effectively restored cell survival compared to Z-VAD-FMK and necrostatin-1s across multiple tumor cell lines (Figure 2B, C). Treatment with ferrostatin-1 increased cell viability in tumor cell lines upon exposure to various concentrations of nintedanib. The restoration of cell survival by Z-VAD-FMK and necrostatin-1s was weaker than ferrostatin-1 in three cell lines (Figure 2D). These results suggested that ferroptosis is partially responsible for the efficacy of nintedanib.

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Nintedanib induces ferroptosis in tumor cells. (A) Representative flow cytometry profiles of annexin V-FITC/PI staining and statistical analysis of MC38 that were pretreated with or without ferrostatin-1 [Ferr-1] (2 μM), carbobenzoxy-valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone [Z-VAD] (5 μM), and necrostatine-1s [Nec-1s] (2 μM) for 24 h followed by treatment with nintedanib (5 μM) for 24 h. (B) Pretreatment with or without ferrostatin-1 [Ferr-1] (2 μM), carbobenzoxy-valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone [Z-VAD] (5 μM), and necrostatine-1s [Nec-1s] (2 μM) for 24 h. A549, HCT116, and MC38 cells were treated with nintedanib (5 μM) for 24 h, then cell viability was detected by CCK8. (C) Clonogenic assay for A549 and MC38 cells that were treated with nintedanib (3 μM) in combination with DMSO, ferrostatin-1 (2 μM), Z-VAD (5 μM), or Nec-1s (2 μM), mean ± SD, n = 3. (D) Cell viability for A549, HCT116, and MC38 cells that were pretreated with or without ferrostatin-1 [Ferr-1] (2 μM), carbobenzoxy-valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone [Z-VAD] (5 μM), and necrostatine-1s [Nec-1s] (2 μM) for 24 h followed by treatment with different dose of nintedanib for 24 h. A549 and MC38 treated with different dose of nintedanib for 24 h. (E) The reactive oxygen species (ROS) level. Scale bar: 20 μm. (F) Lipid peroxidation levels. (G) MDA levels. (H) PTGS2 expression. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Ferroptosis-induced cytotoxicity relies on the production of ROS^40,41. Nintedanib was shown to promote ROS production in tumor cells (Figure 2E). Given that lipid peroxidation is a hallmark of ferroptosis, BODIPY™ 581/591C11 staining was used to assess lipid peroxidation levels. Nintedanib significantly induced lipid peroxidation in tumor cells (Figure 2F). Moreover, MDA, a major byproduct of lipid peroxidation, was also elevated following nintedanib treatment (Figure 2G). Furthermore, the mRNA levels of PTGS2, a key genetic marker of ferroptosis, were upregulated in tumor cells following nintedanib treatment (Figure 2H).

In addition, nintedanib demonstrated a synergistic effect with radiation in promoting ROS production (Figure 3A), inducing lipid peroxidation (Figure 3B) and increasing MDA (Figure 3C) and PTGS2 (Figure 3D) levels. TEM revealed significant morphologic alterations in the cellular ultrastructure of tumor cells treated with nintedanib and radiation with the combined treatment showing the most prominent ferroptotic features, including changes in mitochondrial morphology, such as decreased mitochondrial size, increased membrane density, and reduced or vanished cristae (Figure 3E). Western blot analysis showed that an increase in γH2AX induced by radiation was alleviated by 24 h, whereas nintedanib prolonged γH2AX production, a marker of DNA damage, further supporting the notion that nintedanib enhances tumor cell stress (Figure 3F). These results suggested that nintedanib enhances radiosensitivity by inducing ferroptosis in tumor cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Nintedanib enhances radiosensitivity by promoting ferroptosis in tumor cells. A549 and MC38 cells treated with nintedanib (3 μM) and radiation (8 Gy) for 24 h. (A) The reactive oxygen species (ROS) level. Scale bar: 20 μm. (B) Lipid peroxidation levels. (C) MDA levels. (D) PTGS2 expression. (E) Transmission electron microscopy imaging results of MC38 in different experimental groups. Arrows indicate mitochondria. Scale bar: 500 nm. (F) The expression and quantitative analysis of γH2AX in MC38 cells was determined by western blotting. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Nintedanib enhances ferroptosis sensitivity of tumor cells by suppressing SLC7A11-mediated GSH synthesis

Nintedanib significantly suppressed SLC7A11 expression (Figure 4A), a key regulator of GSH synthesis and utilization. GSH, an antioxidant, has a critical role in preventing lipid peroxidation and ferroptosis⁴². SLC7A11 expression was reduced by radiation in the MC38 cell line, while SLC7A11 expression was activated in the A549 cell line (Figure 4B). Despite these different responses to radiation, nintedanib consistently inhibited SLC7A11 expression in both cell lines, leading to a reduction in GSH levels (Figure 4C, D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Nintedanib induces ferroptosis through inhibiting SLC7A11-mediated GSH synthesis. (A, B) The expression and quantitative analysis of GAPDH and SLC7A11 in A549 and MC38 cells was determined by western blotting. (C, D) GSH levels in A549 and MC38 cells after different treatments. (E, F) GSH levels and cell viability for A549 cells that were pretreated with or without N-acetyl cysteine [NAC] (5 mM) for 24 h followed by treatment with nintedanib for 24 h. (G) The expression of SLC7A11 in A549 cells expressing vector control (NC) or SLC7A11 was determined by western blotting. (H–J) Cell viability, GSH levels, and lipid peroxidation levels in NC and OE-SLC7A11 A549 cells after different treatments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

NAC was used to promote GSH production to further explore the relationship between GSH synthesis and nintedanib-induced ferroptosis. NAC significantly enhanced GSH levels in nintedanib-treated and untreated A549 cells (Figure 4E) and restored cell viability following nintedanib treatment (Figure 4F). In addition, SLC7A11-overexpressing A549 cells were generated to confirm the role of SLC7A11 in nintedanib-induced ferroptosis (Figure 4G). Overexpression of SLC7A11 notably reversed the reduction in cell viability (Figure 4H), GSH depletion (Figure 4I), and lipid peroxidation enhancement (Figure 4J) induced by nintedanib and radiation.

These results indicated that nintedanib sensitizes tumor cells to radiation by inhibiting SLC7A11-mediated GSH synthesis.

ATF4 is involved in nintedanib-induced SLC7A11 downregulation

Previous studies identified stress sensors, including ATF4, NRF2, ATF3, and P53, as regulators of SLC7A11 expression^43–45. To assess the impact of nintedanib on these regulatory proteins, expression was evaluated following treatment. Western blot analysis revealed that nintedanib significantly inhibited ATF4 expression, while having no effect on other SLC7A11 regulators (Figure 5A). Given that ATF4 binds to amino acid response elements to promote SLC7A11 transcription^46–48, it was hypothesized that ATF4 is a crucial target through which nintedanib induces ferroptosis in tumor cells. RNA-seq analysis of downstream genes of ATF4 in nintedanib-treated tumor cells revealed a gene expression profile resembling that of ATF4-deficient hepatocytes (Figure 5B), indicating increased sensitivity to ferroptosis⁴⁹. This further supports the idea that nintedanib induces ferroptosis by inhibiting ATF4 expression.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

ATF4 knockdown promoted radiosensitivity of tumor cells. (A) The expression and quantitative analysis of GAPDH, ATF4, ATF3, NRF2, and P53 in A549 and MC38 cells was determined by western blotting. (B) The level of ATF4-related genes in RNA-seq data of MC38 cells. (C) The expression and quantitative analysis of GAPDH, ATF4, and SLC7A11 in MC38, A549, and HCT116 treated with increasing dose of radiation. (D, E) PTGS2 expression and lipid peroxidation level in A549 (shNC and shATF4) and HCT116 (shNC and shATF4) after different treatments for 24 h. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

In addition, the effect of radiation on ATF4 expression varied among different cell types. Radiation led to downregulation of ATF4/SLC7A11 in the MC38 cell line and upregulation in A549 and HCT116 cells (Figure 5C). This finding suggests that ATF4 may contribute to radio-resistance by preventing ferroptosis in tumor cells. To investigate this finding further, ATF4 knockdown and ATF4 overexpression cell lines were created in A549 and HCT116 cells (Figure S4A and S4C). Colony formation assays demonstrated that ATF4 has a role in radiation resistance (Figure S4B and S4D). Elevated levels of PTGS2 and lipid peroxidation in ATF4 knockdown cells indicated that ATF4 mediates resistance to radiation-induced ferroptosis (Figure 5D, E). Furthermore, ATF4 expression was correlated with a poor prognosis in lung cancer (GSE30219, GSE37745, and GSE4573; Figure S4E) and colorectal cancer (GSE17538; Figure S4E) as revealed by Kaplan–Meier curves (https://kmplot.com). These results suggest the potential of targeting ATF4 to enhance the antitumor efficacy of radiotherapy.

Nintedanib was shown to reverse the radiation-induced upregulation of ATF4 expression in A549 and HCT116 cell lines (Figure 6A). ATF4-overexpressing A549 and MC38 cell lines were constructed to confirm the effect of nintedanib on ATF4 (Figure 6B). ATF4 overexpression mitigated the reduction in cell viability (Figure 6C), the increase in PTGS2 transcription (Figure 6D), and the enhancement of lipid peroxidation (Figure 6E) caused by nintedanib and radiation. Furthermore, subcutaneous tumor models were established by injecting these cell lines into the flanks of nude mice. Treatment with nintedanib and RT significantly inhibited tumor growth, an effect that was partially reversed by ATF4 overexpression (Figure 6F, G).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Nintedanib enhances radiosensitivity of tumor cells by inhibiting ATF4. (A) The expression and quantitative analysis of GAPDH and ATF4 in A549 and HCT116 cells was determined by western blotting. A549 and MC38 cells were transfected with an empty vector or ATF4 plasmids. (B) The expression of GAPDH and ATF4 in A549 and MC38 cells expressing vector control (NC) or ATF4 was determined by western blotting. (C–E) Cell viabilities, PTGS2 expression and lipid peroxidation levels in A549 (NC and OE-ATF4) and MC38 (NC and OE-ATF4) after different treatments for 24 h. MC38 cells (vector control and OE-ATF4) were implanted subcutaneously into C57BL/6 mice (n = 7 per group). (F, G) Tumor growth and survival curves of tumor-bearing mice (n = 7 per group). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

CD31, an angiogenic biomarker, is commonly used to assess blood vessel density in tumor tissues, while α-SMA⁺/CD31⁺ expression reflects tumor vascular perfusion, indicating vascular normalization⁵⁰. Immunofluorescence staining for CD31 and α-SMA was performed on tumor tissues to determine whether the angiogenic effects of nintedanib are linked to ATF4 inhibition. The results (Figure S5A–C) demonstrated that the nintedanib anti-angiogenic effect was independent of ATF4 inhibition.

Nintedanib combined with RT enhances α-PD-L1 therapy efficacy

Recent studies have highlighted that ferroptosis-related lipid peroxides act as identifying signals that promote the recognition and processing of tumor antigens by dendritic cells, activate cytotoxic T lymphocytes, and enhance tumor immunotherapy⁵¹. In addition, the efficacy of immunotherapy has long been considered to be associated with PD-L1 expression in tumors⁵². Therefore, PD-L1 expression was determined on the surface of tumor cells after nintedanib combined with radiation stimulation. The combined treatment further upregulated PD-L1 in tumors compared to other treatments, indicating the potential of the therapeutic strategy combining nintedanib with RT and anti-PD-L1 therapy, as shown in Figure 7A. Next, animal models were used to verify that the triple treatment strategy further inhibited tumor volume growth and prolonged the tumor-bearing survival time of mice compared to the dual treatment strategy (Figure 7B–E). Immunohistochemistry was used to determine whether the number of CD8⁺ cells in the tumor center differed in each group. As demonstrated by the positive cell count analysis, triple therapy group tumors had a greater CD8⁺ cell infiltrate compared to the other treatment groups (Figure 7F). Furthermore, flow cytometry analysis revealed that the proliferation of CD3⁺ and CD8⁺ T cells was most robust in the combination treatment group (Figure 7G–I). Therefore, nintedanib combined with RT promotes ferroptosis in tumors and enhances the effect of immunotherapy by up-regulating PDL1 in tumors and increasing the number and function of tumor-infiltrating T cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

The antitumor activity of nintedanib combined with radiotherapy and immunotherapy. (A) LLC cells treated with nintedanib (3 μM) and radiation (8 Gy) for 24 h. Surface expression of PD-L1 were determined by flow cytometry. (B) Treatment schedule of tumor-bearing mice. LLC cells were implanted subcutaneously into C57BL/6 mice (n = 7 per group). (C, D) Tumor growth curves and survival curves of LLC tumor-bearing mice. (E) Individual tumor volume curves in LLC tumor survival experiment of C57BL/C mice. (F) Representative images of CD8 IHC staining in LLC tumors and their quantitative analysis. Scale bar: 50 μm. (G–I) Ki-67 in CD3⁺ T cells and CD3⁺CD8⁺ T cells from TIL of LLC tumors and their quantitative analysis. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Preliminary biosafety evaluations of combination treatment of nintedanib and RT in a mouse model

Finally, the safety of nintedanib in combination with RT was assessed. Mouse weight monitoring throughout the experiment did not show significant changes (Figure S6A, B). Hematoxylin and eosin staining of major organs revealed no pathologic alterations (Figure S6C) and routine blood biochemical analyses indicated no significant abnormalities in organ function (Figure S6D). No deterioration of the overall health status of mice, significant weight loss, or spontaneous death of the animals was observed during the experiments. These results provide preliminary evidence supporting the efficacy and safety of nintedanib combined with RT in tumor treatment.