Exogenous iron impairs the anti-cancer effect of ascorbic acid both in vitro and in vivo

Graphical abstract

Highlights

• •Ascorbic acid triggered ROS/Ca²⁺ dependent necrosis.
• •Extracellular Fe²⁺/Fe³⁺ abrogated the cytotoxicity of ascorbic acid in vitro.
• •Fe²⁺/Fe³⁺ supplement diminished the anti-cancer effects of ascorbic acid in vivo.

Abstract

Introduction

The anti-cancer effect of high concentrations of ascorbic acid (AA) has been well established while its underlying mechanisms remain unclear. The association between iron and AA has attracted great attention but was still controversial due to the complicated roles of iron in tumors.

Objectives

Our study aims to explore the anti-cancer mechanisms of AA and the interaction between AA and iron in cancer.

Methods

The MTT and ATP assays were used to evaluate the cytotoxicity of AA. Reactive oxygen species (ROS) generation, calcium (Ca²⁺), and lipid peroxidation were monitored with flow cytometry. Mitochondrial dysfunction was assessed by mitochondrial membrane potential (MMP) detection with JC-1 or tetramethylrhodamine methyl ester (TMRM) staining. Mitochondrial swelling was monitored with MitoTracker Green probe. FeSO₄ (Fe²⁺), FeCl₃ (Fe³⁺), Ferric ammonium citrate (Fe³⁺), hemin chloride (Fe³⁺) were used as an iron donor to investigate the effects of iron on AA’s anti-tumor activity. The in vivo effects of AA and iron were analyzed in xenograft zebrafish and allograft mouse models.

Results

High concentrations of AA exhibited cytotoxicity in a panel of cancer cells. AA triggered ROS-dependent non-apoptotic cell death. AA-induced cell death was essentially mediated by the accumulated intracellular Ca²⁺, which was partly originated from endoplasmic reticulum (ER). Surprisingly, exogenous iron could significantly reverse AA-induced ROS generation, Ca²⁺ overloaded, and cell death. Especially, the iron supplements significantly impaired the in vivo anti-tumor activity of AA.

Conclusions

Our study elucidated the protective roles of iron in ROS/Ca²⁺ mediated necrosis triggered by AA both in vitro and in vivo, which might shed novel insight into the anti-cancer mechanisms and provide clinical application strategies for AA in cancer treatment.

Introduction

It has sparked a lot of debate since the anti-cancer potential of ascorbic acid (AA) was put forward more than 50 years ago [1]. Pharmacological concentrations (mM) of AA are confirmed to effectively kill various cancer cells, leaving most of the normal cells unaffected [2], [3]. Various underlying mechanisms have been proposed in recent years (e.g., redox imbalance, epigenetic regulation, HIF signaling regulation, glucose metabolism) [4], [5], [6]. However, the characterizations of AA-induced cell death remain unclear.

Iron is essential for various physiological processes including DNA synthesis, electron transport, oxygen transport, etc., and is strictly regulated under normal physiological conditions [7], [8]. A growing number of studies have uncovered the association between iron and cancer [9]. Ferroptosis, a type of programmed cell death that is tightly regulated by intracellular iron and iron-dependent lipid peroxidation, has aroused great interest in cancer therapy [10]. As a redox-active transition metal, iron has been identified to enhance the oxidative stress induced by AA [11], [12]. AA was also identified to regulate iron metabolism in cancer cells [13]. However, the role of iron on the anti-cancer activity of AA remains unclear and several in vitro studies reported inconsistent observations. Carosio et al [14] reported that intracellular iron depletion contributes to AA-induced apoptosis in neuroblastoma cell lines. Tsuma-Kaneko et al [15] found that AA-induced growth inhibition and apoptosis in K562 leukemic cells were enhanced by iron removal. While Mojić et al [16] reported that the anti-cancer effects of AA in PC-3 and LNCaP prostate cancer cell lines were diminished by extracellular iron. To better explore the anti-cancer mechanisms of AA, this study focused on AA-mediated cell death and the effect of iron on this process using in vitro and in vivo models.

Materials and Methods

Reagents

(+)-Sodium L-ascorbate (AA, ≥ 98%), deferoxamine mesylate salt (DFO, ≥ 92.5%), 3-methyladenine (3-MA, ≥ 98%), ferrostatin-1 (Fer-1, ≥ 95%), chloroquine (CQ, ≥ 98%), ethylenebis(oxyethylenenitrilo)tetraacetic acid (EGTA,  ≥ 98%), and 2-APB (≥ 98%) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Necrostatin-1 s (Nec-1 s, ≥ 96%) was obtained from BioVision (Milpitas, CA, USA). Glycine ( ≥ 99%), catalase, and N-acetyl-l-cysteine (NAC,  ≥ 99%) were obtained from Beyotime (Shanghai, China). N-benzyloxycarbonyl-Val-Ala-Aspfluoromethylketone (Z-VAD-FMK,  ≥ 99%), GSK’872 (≥ 99%), VX765 (≥ 99%), and azoramide (≥ 99%) were obtained from Selleck Chemicals (Houston, TX, USA). Ferric ammonium citrate (FAC,  ≥ 98%), FeSO₄ (≥ 98%), and FeCl₃ (≥ 98%) were obtained from Macklin (Shanghai, China). BAPTA-AM (≥ 98%) was obtained from Molecular Probes (Eugene, OR, USA). Hemin chloride (Hemin,  ≥ 95%) was obtained from J&K Scientific Ltd. (Beijing, China). 4-Phenylbutyrate (4-PBA,  ≥ 98%) was obtained from Aladdin (Shanghai, China).

Cell lines and cell culture

Cancer cell lines were obtained from American Type Culture Collection (Manassas, VA, USA). A549, NCI-H1975, NCI-H460, NCI-H1299 human lung cancer cell lines, and murine colon carcinoma cell line CT26, murine mammary carcinoma cell line 4T1, were cultured in 10% fetal bovine serum-containing RPMI 1640. Human colon cancer cell line HCT116 was cultured in 10% fetal bovine serum-containing McCoy's 5A. Cells were maintained in an incubator at 37°C with 5% CO₂.

MTT assay

3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide (MTT) reagent was used for cell viability determination as previous report [17]. The absorbance was measured with FlexStation 3 microplate reader.

Intracellular ATP determination

Intracellular ATP level was evaluated with CellTiter-Glo luminescent assay reagents (Promega) as manufacturer’s instructions. The luminescent signal was measured with FlexStation 3 microplate reader.

Intracellular ROS measurement

Intracellular ROS was measured using 2′,7′-dichlorofluorescin diacetate (H₂DCFDA) (Invitrogen) as previous report [18]. Cells that received AA or other combination treatments were loaded with H₂DCFDA (10 μM), followed by flow cytometric analysis using BD LSRFortessa™ Cell Analyzer (BD Bioscience).

Mitochondrial ROS measurement

Mitochondrial ROS was measured using MitoSOX Red (Invitrogen) following the manufacturer’s instructions. Cells that received AA or other combination treatments were loaded with MitoSOX Red (5 μM), followed by flow cytometric analysis using BD LSRFortessa™ Cell Analyzer.

Lipid peroxidation assay

BODIPY™ 581/591 C11 probe (Invitrogen) was used for lipid peroxidation determination. Cells that received AA or other combination treatments were loaded with BODIPY (5 μM), followed by flow cytometric analysis using BD LSRFortessa™ Cell Analyzer.

Annexin V/7AAD double staining

Apoptosis was evaluated with PE Annexin V Apoptosis Detection Kit I (BD Bioscience) following the manufacturer’s instructions. Cisplatin served as a positive control.

Western blotting

Cells that received AA or other treatments were lysated with protease inhibitor- and phosphatase inhibitor-containing RIPA lysis buffer. The bicinchoninic acid assay was used for protein content determination. 30 μg protein samples were separated using SDS-PAGE and electrotransferred onto polyvinylidene difluoride membrane, which were then blocked with non-fat milk in TBST buffer (2 h) before incubation with primary antibody (4°C, 12 h) and secondary antibody (37°C, 2 h). The ChemiDoc™ MP Imaging System was used for chemiluminescence detection.

Propidium iodide (PI) staining

The cell membrane integrity was indicated with PI staining as previous report [17]. Cells that received AA or other combination treatments were stained with PI (50 μg/mL). The IncuCyte® S3 Live-Cell Analysis System was used for fluorescence image acquisition.

Intracellular Ca²⁺ measurement

Intracellular Ca²⁺ was measured with Fluo-3 AM (Invitrogen) as the manufacturer's protocols. Cells that received AA or other combination treatments were loaded with Fluo-3 AM (5 μM), followed by flow cytometric analysis using BD LSRFortessa™ Cell Analyzer (BD Bioscience).

Mitochondrial Ca²⁺ measurement

Mitochondrial Ca²⁺ was measured with Rhod-2 AM (Invitrogen) as the manufacturer's protocol. Cells that received AA or other combination treatments were loaded with Rhod-2 AM (5 μM), followed by flow cytometric analysis using BD LSRFortessa™ Cell Analyzer.

JC-1 staining

The mitochondrial membrane potential (MMP) was measured with JC-1 staining kit (Beyotime). Cells that received AA or other combination treatments were loaded with JC-1 following the manufacturer’s instructions.

Tetramethylrhodamine methyl ester (TMRM) staining

The MMP was measured with TMRM staining (Invitrogen). Cells that received AA or other combination treatments were loaded with TMRM (100 nM), followed by fluorescence microscopy imaging.

MitoTracker Green staining

The mitochondria were tracked with MitoTracker Green (Invitrogen). Cells that received AA or other combination treatments were loaded with MitoTracker Green (100 nM), followed by confocal fluorescence microscopy imaging.

Intracellular iron measurement

Intracellular iron was measured with Calcein AM (Invitrogen) as the manufacturer's protocols. Cells received AA or other combination treatments were loaded with Calcein AM (20 nM), followed by flow cytometric analysis using BD LSRFortessa™ Cell Analyzer.

Zebrafish xenograft model

The establishment of the zebrafish A549-derived xenograft model has been previously described [17]. Briefly, 48 h post-fertilization zebrafish larvae were collected. Approximate 200 A549 cells labeled with DiI (Invitrogen) were transplanted into the yolk sac. After 3 h, zebrafish were divided into 8 groups: Saline; AA (2 mM); FeSO₄ (100 μM); FAC (100 μM); hemin (100 μM); AA (2 mM) + FeSO₄ (100 μM); AA (2 mM) + FAC (100 μM); AA (2 mM) + hemin (100 μM). After 72 h exposure to drug-containing fish water, fluorescence images were acquired with SMZ18 microscope. All the fish water contained 1-phenyl-2-thiourea (0.2 mM) to suppress pigmentation.

Mouse allograft model

8-week male BALB/c mice were subcutaneously implanted with CT26 (5 × 10⁵) cells on the right lower flank and then divided into 4 groups when tumors reached 100 mm³ (calculated as 1/2 × lengths × widths²) and received intraperitoneal injection daily: Saline; AA (4 g/kg); AA (4 g/kg) + FAC (20 mg/kg); AA (4 g/kg) + hemin (20 mg/kg). After 14 days, the mice were sacrificed.

Ethics statement

Animal experiments were conducted following the ethical guidelines approved by the Animal Research Ethics Committee of the University of Macau (Ethics ID: UMARE-016–2021, UMARE-004–2019).

Malondialdehyde (MDA) measurement

MDA levels in tumor tissues were determined with the MDA detecting kit as manufacturer's protocols.

Statistical analysis

Data were processed using GraphPad Prism 8.0 and showed as means ± standard deviation (SD). Statistical significance was analyzed by Student's t-test or one-way ANOVA with Dunnett's post hoc test. p < 0.05 represented significant.

Results

AA triggered ROS-mediated cell death in cancer cells

The cytotoxicity of AA was evaluated in several cancer line cells. As shown in the MTT and ATP assays, AA dose- and time-dependently induced cell death in cancer cells (Fig. 1A, Supplementary Fig. 1). AA time-dependently induced intracellular ROS and mitochondrial ROS (mROS) generation (Fig. 1B). Furthermore, AA induced a significant mitochondrial membrane potential (MMP) depletion and mitochondrial swelling (Fig. 1C). Increased lipid peroxidation was also detected (Fig. 1D). AA-induced intracellular ROS, mROS, lipid peroxidation, and MMP depletion were significantly reversed by the ROS scavengers NAC or catalase (Fig. 1E-H). Furthermore, both ROS scavengers nearly completely rescued cell death induced by AA as determined in MTT and ATP assays (Fig. 1I). These results indicated that mM levels of AA triggered ROS-mediated cancer cell death.

Fig. 1.

AA triggered ROS-mediated cell death. (A) A549 cells were treated with AA (2 mM, 4 h), then washed with PBS, the MTT (left), or ATP (right) assay was performed after 2, 8, 20 h. (B) A549 cells were treated with AA (2 mM) for indicated times, and intracellular ROS (left) or mROS (right) generation was detected with flow cytometry. TBHP served as the positive control. (C) A549 cells were treated with AA (2 mM, 2 h), and MMP was detected with JC-1 and TMRM staining, mitochondria were tracked with MitoTracker Green. (D) A549 cells were treated with AA (2 mM) for indicated times, and lipid peroxidation was detected with flow cytometry and fluorescence microscopy. (E-G) A549 cells were treated with AA (2 mM, 2 h) with or without pretreatment of catalase (2000 U) or NAC (5 mM), and (E) intracellular ROS, (F) mROS generation, or (G) lipid peroxidation was detected with flow cytometry. (H) A549 cells were treated with AA (2 mM, 2 h) with or without pretreatment of NAC (5 mM). And MMP was detected with JC-1. (I) A549 cells were treated with AA (2 mM, 4 h) with or without pretreatment of, then washed with PBS. After 20 h, the MTT (left), or ATP (right) assay was performed. Bar = 10 μm. ^**p < 0.01. ^***p < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

AA-triggered non-apoptotic cell death

We then examined whether apoptosis was involved in AA-induced cell death. As shown in Fig. 2A, the pan-caspase inhibitor Z-VAD-FMK failed to rescue the cytotoxicity of AA in A549 cells. Neither Annexin V(+)/7AAD(-) cells nor cleavage of caspase 3, 7 in Western blotting was detected after AA treatment (Fig. 2B, 2C). Plenty of PI-positive cells were observed after AA treatment, indicating the destruction of cell membrane integrity (Fig. 2D). Furthermore, the inhibitors of necroptosis (Nec-1s, GSK’872), autophagy (3-MA, CQ), ferroptosis (Fer-1, DFO), pyroptosis (VX765, glycine) showed no reversal effect on AA-triggered cell death (Fig. 2E). AA treatment showed no effect on the expression of key proteins that mediated the cell death (Fig. 2F). Taken together, these results revealed that AA triggered non-apoptotic but necrotic cell death.

Fig. 2.

AA-triggered non-apoptotic cell death. (A) A549 cells were treated with AA (2 mM, 4 h) with or without pretreatment of Z-VAD-FMK, then washed with PBS. The MTT assay was performed after 20 h. (B) A549 cells were treated with AA (2 mM, 4 h), and apoptosis was detected by Annexin V/7AAD staining. Cisplatin (CIS) (0.3 mM, 24 h) served as the positive control. (C) A549 cells were treated with AA (2 mM) for indicated times and cleaved caspase 3/7 (C-Cas 3/7) were detected. CIS (0.3 mM, 24 h) served as the positive control. (D) The cell membrane integrity of A549 cells was detected with PI staining after AA treatment. Bar = 100 μm. (E) A549 cells were treated with AA (2 mM, 4 h) with or without pretreatment of Nec-1s, GSK’872, CQ, 3-MA, DFO, Fer-1, VX765, glycine, respectively, then washed with PBS. The MTT assay was performed after 20 h. (F) A549 cells were treated with AA (2 mM) for indicated times, and P-RIP1/3, RIP1/3, p62, LC3-I/II, GPX4, CYPOR, GSDMD were detected. β-actin served as the loading control. ^**p < 0.01. ^***p < 0.001.

AA induced Ca²⁺-mediated cell death

Notably, AA-induced significant increase in both intracellular and mitochondrial Ca²⁺ (Fig. 3A). AA-induced cell death was remarkably inhibited by the intracellular Ca²⁺ chelator BAPTA-AM but not the extracellular Ca²⁺ chelator EGTA (Fig. 3B), indicating that AA-induced cell death was mediated by endogenous Ca²⁺. The inositol 1,4,5-triphosphate receptor (IP₃R) inhibitor 2APB could partially reverse AA-induced accumulation of both intracellular and mitochondrial Ca²⁺ (Fig. 3C), indicating that Ca²⁺ was partially released from ER. 2APB only alleviated AA-induced ATP depletion in 2 h but failed to do so in 24 h. While BAPTA-AM could significantly inhibit AA-induced ATP-depletion even at 24 h (Fig. 3D). Furthermore, AA induced time-dependent phosphorylation of eIF2α suggesting the ER stress. However, the ER stress inhibitors azoramide and 4-PBA showed no effect on AA-induced cell death (Fig. 3E, 3F). The ROS scavengers (NAC and catalase) and BAPTA-AM significantly reversed AA-induced eIF2α phosphorylation and Ca²⁺ overload (Fig. 3G, 3H). Collectively, these results revealed that AA-induced cell death was mediated by Ca²⁺.

Fig. 3.

AA induced Ca²⁺-mediated cell death. (A) A549 cells were treated with AA (2 mM) for indicated times, then intracellular (left) or mitochondrial (right) Ca²⁺ were detected with flow cytometry. (B) A549 cells were treated with AA (2 mM, 4 h) with or without pretreatment of BAPTA-AM or EGTA, then washed with PBS. After 20 h, the MTT (left) or ATP (right) assay was performed. (C) A549 cells were treated with AA (2 mM, 2 h) with or without pretreatment of 2APB (100 μM), and intracellular (left) or mitochondrial (right) Ca²⁺ was detected with flow cytometry. (D) A549 cells were treated with AA (2 mM) for 2 h (left) or 24 h (right) with or without pretreatment of BAPTA-AM or 2APB, and the ATP level was detected. (E) A549 cells were treated with AA (2 mM) for indicated times and P-eIF2α, eIF2α were detected. (F) A549 cells were treated with AA (2 mM, 4 h) with or without pretreatment of azoramide or 4-PBA, then washed with PBS. After 20 h, the MTT or ATP assay was performed. (G) A549 cells were treated with AA (2 mM, 2 h) with or without pretreatment of NAC (5 mM), catalase (2000 U), BAPTA-AM (2 μM), or 2APB (100 μM), respectively, and P-eIF2α, eIF2α were detected. (H) A549 cells were treated with AA (2 mM, 2 h) with or without pretreatment of NAC (5 mM) or catalase (2000 U), and intracellular (left) or mitochondrial (right) Ca²⁺ level was detected with flow cytometry. β-actin served as the loading control. ^**p < 0.01. ^***p < 0.001.

Iron protected AA-induced cell death and mitochondrial dysfunction

Recent studies have gradually shed light on the role of iron in various physiological processes including cancer. To explore whether iron was also involved in regulating AA-triggered cancer cell death, the intracellular iron level was evaluated. Results demonstrated that AA treatment decreased intracellular iron levels (Supplementary Fig. 2A). Interestingly, Fe²⁺ (FeSO₄) and Fe³⁺ (FeCl₃, FAC, hemin) nearly completely abolished the cytotoxic effect of AA in A549 cells (Fig. 4A, 4B). The reversal effects of Fe²⁺/Fe³⁺ were also observed in other types of cancer cell lines including NCI-H1975, NCI-H460, NCI-H1299, HCT116, CT26, and 4T1 (Supplementary Fig. 3). However, other cations including Cu⁺, Cu²⁺, Mg²⁺, Zn²⁺, and Al³⁺ failed to do so (Fig. 4C). Furthermore, Fe²⁺/Fe³⁺ remarkably attenuated AA-induced ROS generation, Ca²⁺ overload, as well as lipid peroxidation (Fig. 4D-G). Meanwhile, AA-induced mROS, mitochondrial Ca²⁺, MMP depletion, and mitochondrial swelling were nearly completely reversed by Fe²⁺/Fe³⁺ (Fig. 5A-D). Further, AA-induced ER stress were abrogated by Fe²⁺/Fe³⁺ (Fig. 5E). These results revealed the critical role of iron in AA-triggered cell death.

Fig. 4.

Iron protected AA-induced cell death. (A-B) A549 cells were treated with AA (2 mM, 4 h) with or without pretreatment of FeSO₄, FeCl₃, FAC, or hemin, respectively, then washed with PBS. After 20 h, the MTT assay was performed, and the morphological changes were acquired with a microscope. (C) A549 cells were treated with AA (2 mM, 4 h) with or without pretreatment of CuCl, CuCl₂, MgSO₄, ZnCl₂, or AlCl₃, respectively, then washed with PBS. The MTT assay was performed after 20 h. (D-G) A549 cells were treated with AA (2 mM, 2 h) with or without pretreatment of FeSO₄, FeCl₃, FAC, or hemin, respectively, and (D) ROS generation, (E) Ca²⁺ level, or (F) lipid peroxidation was detected with flow cytometry. (G) Lipid peroxidation was also detected with a fluorescence microscope. Bar = 100 μm. ^**p < 0.01. ^***p < 0.001.

Fig. 5.

Iron protected AA-induced mitochondrial dysfunction and ER stress. (A-D) A549 cells were treated with AA (2 mM, 2 h) with or without pretreatment of FeSO₄, FeCl₃, FAC, or hemin, respectively. Then (A) mROS and (B) mitochondria Ca²⁺ were detected with flow cytometry, (C) MMP was evaluated with JC-1 staining, (Bar = 100 μm), (D) mitochondria swelling was monitored with MitoTracker Green. (Bar = 10 μm). (E) A549 cells were treated with AA (2 mM, 2 h) with or without pretreatment of FeSO₄, FeCl₃, FAC, or hemin, respectively, and P-eIF2α, eIF2α were detected. β-actin served as the loading control. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Iron impaired the in vivo anti-tumor activity of AA

The effect of iron on AA’s anti-tumor activity was further investigated in vivo. In the A549-derived zebrafish xenograft model, AA significantly suppressed tumor growth, which was efficiently reversed by the combination treatment of FeSO₄, FAC, or hemin (Fig. 6A). Furthermore, AA reduced tumor growth and weights in the CT26-derived mouse allograft model, which were partially reversed by FAC and hemin as well (Fig. 6B-D). In addition, AA treatment increased tumor tissue MDA levels, which were reduced by FAC and hemin (Fig. 6E). These results demonstrated that iron impaired the in vivo anti-tumor activity of AA, which was consistent with the in vitro results.

Fig. 6.

Iron diminished the in vivo anti-tumor activity of AA. (A) The zebrafishes transplanted with cancer cells were exposed to AA (2 mM) alone or in combination with FeSO₄, FAC, hemin (100 μM) for 3 days. The in vivo tumor fluorescence images were acquired (left) and fluorescence area was calculated (right) (n = 6). (B) The tumor tissues were retrieved from treated mice. (C-D) The average tumor volume and final tumor weight. (E) The tumor tissue MDA levels. ^**p < 0.01. ^***p < 0.001.

Discussion

The anti-cancer mechanisms of of AA remain controversial and may involve multiple signaling pathways [19]. The main findings of the current study are: a). AA triggered ROS/Ca²⁺-mediated non-apoptotic necrosis. b). The anti-cancer efficacy of AA was impaired by exogenous iron both in vitro and in vivo.

The anti-cancer mechanisms of AA have been extensively studied. Previous studies demonstrated that AA induces different types of cell death including necroptosis, autophagy, apoptosis, e.g., depending on the concentration and the cell type [20]. Redox imbalance, oxygen-sensing regulation and epigenetic reprogramming were the three vulnerabilities of many cancer cells that could be targeted by AA [4]. Consist with previous reports [20], [21], we showed that AA triggered ROS generation and lipid peroxidation in A549 cells, which were attenuated by NAC and catalase. Meanwhile, the JC-1, TMRM, and MitoTracker Green staining showed that AA-induced significant mitochondrial dysfunction, which was reversed by ROS scavengers as well. These results confirmed the critical role of oxidative stress in AA-triggered cell death.

However, Annexin V/7-AAD double staining detected few apoptotic cells and no morphological apoptotic characteristics were observed in electron microscope assay (data not shown). Furthermore, the apoptosis-related caspases inhibition by Z-VAD-FMK failed to alleviate AA-induced cell death and no cleaved-caspase 3/7 was detected after AA treatment. On the contrary, PI staining indicated that AA-induced cell membrane rupture, a characteristic of necrosis. Thus, AA-triggered non-apoptotic necrosis. In addition, inhibitors of necroptosis, autophagy, ferroptosis or pyroptosis failed to protect against cell death induced by AA. The key biomarker proteins of these cell death types were not affected by AA. Collectively, these results suggested that AA induces an unidentified type of necrosis.

Ca²⁺ is a critical mediator in different cell death including necrosis, apoptosis, anoikis, autophagy, etc. [22], [23]. A recent report showed that AA triggered necrosis in human laryngeal squamous cells through ROS, PKC, and Ca²⁺ signaling [24]. Similar Ca²⁺ overload was observed in AA-treated A549 cells. BAPTA-AM, the intracellular Ca²⁺ chelator, significantly reversed AA-induced ROS generation and cell death while EGTA, the extracellular Ca²⁺ chelator, failed to do so. Thus, the AA-induced cell death may be due to the release of intracellular Ca²⁺ stores. Previous studies reported the role of PKC activation [24] and phospholipase C in mediating AA-induced Ca²⁺, necrosis, and apoptosis [25]. The intracellular Ca²⁺ homeostasis was tightly regulated via diverse signalings and ER was one of the main internal Ca²⁺ stores [26], [27]. Furthermore, a previous study showed that Ca²⁺ release potentiates the ER stress and cell death induced by AA-driven menadione redox cycling [28]. AA triggered eIF2 phosphorylation indicating ER stress. The blockage of ER Ca²⁺-release receptor channel IP₃R with 2APB partially reversed AA-induced Ca²⁺ overload. It transiently attenuated AA-induced ATP depletion and failed to reverse AA-induced cell death as BAPTA-AM did. Thus, AA-induced cell death was mediated by intracellular Ca²⁺ increase but IP₃R-mediated Ca²⁺ release from ER may have a limited effect in this process. Given the complicated regulatory system of intracellular Ca²⁺, the precise source of Ca²⁺ in response to AA remains to clarify. In addition, ER stress inhibitors showed no effect on AA-induced cell death while ROS scavengers reversed both intracellular and mitochondrial Ca²⁺ overload, and ER stress. Thus, ROS was an upstream modulator of Ca²⁺ overload and ER stress in AA-treated cells.

Iron, a transition metal, plays important roles in various physiological and pathological processes while its roles in cancer remain controversial [9]. Ferroptosis, a recently identified non-apoptotic programmed cell death tightly regulated by iron, shows therapeutic potential in cancer [10]. Some reports suggested that AA-triggered cell death involves ferroptosis [20] while other studies showed that iron removal enhances AA-induced apoptosis [14], [29]. Especially, extracellular FAC diminished the anticancer effects of AA in prostate cancer cell lines [16]. Interestingly, we found that the cytotoxic effect of AA was significantly reversed by both Fe³⁺ (FAC, FeCl₃, hemin) and Fe²⁺ (FeSO₄) but not other cations (Cu⁺, Cu²⁺, Mg²⁺, Zn²⁺, Al³⁺). This phenomenon was observed in a panel of cancer cell lines. Thus, it indicated that the protective effect was iron specific and might be a universal effect (Supplementary Fig. 3). Consist with this, we found that AA treatment decreased intracellular free iron levels (Supplementary Fig. 2A) and ferritin heavy chain 1 (FTH1), a major regulator of intracellular Fe²⁺/Fe³⁺[30], was increased within 1 h after AA treatment (Supplementary Fig. 2B). A study revealed that high concentrations of AA produce ascorbate radical and extracellular H₂O₂. The latter diffused into cells, depleted ATP, and induced cell death [31]. Our results supported this as we observed that pre- and co-treatment of iron donors inhibited AA’s cytotoxicity while post-treatment showed no effects (data not shown). It suggested that extracellular ROS generated by AA might be quickly decomposed by iron donors before their diffusion into the cells. The critical role of iron in the AA-triggered anticancer effect was further confirmed by in vivo results. AA showed a significant anticancer effect in both tumor-bearing zebrafish and mouse models. The significant reversal effects of FAC, FeSO₄, and hemin revealed that iron suppelementation impaired AA’s anti-cancer effects.

In conclusion, our data showed that AA triggered ROS/Ca²⁺-dependent necrosis which was compromised by iron in vitro and in vivo, suggestting that iron supplementation may decrease the efficacy of AA in clinical trials.

CRediT authorship contribution statement

Bingling Zhong: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing – original draft, Visualization. Lin Zhao: Methodology, Investigation. Jie Yu: Methodology. Ying Hou: Methodology, Investigation. Nana Ai: Methodology. Jin-Jian Lu: Writing – review & editing. Wei Ge: Resources, Writing – review & editing. Xiuping Chen: Resources, Writing – review & editing, Supervision, Project administration, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

The following are the Supplementary data to this article: