Arsenic trioxide induces ferroptosis in neuroblastoma by mediating GPX4 transcriptional inhibition

Abstract

Neuroblastoma (NB), the most common extracranial solid tumor in childhood, significantly contributes to cancer‐related mortality, presenting a dearth of efficacious treatment strategies. Previously, our studies have substantiated the potent cytotoxicity of arsenic trioxide (ATO) against NB cells, however, the specific underlying mechanism remains elusive. Here, we first identified ATO as a novel GPX4 inhibitor, which could trigger the ferroptosis in NB cells. In vitro, ATO significantly inhibited the proliferation and migration ability of NB cells SK‐N‐AS and SH‐SY5Y, and induced ferroptosis. Furthermore, the iron chelator deferoxamine reversed ATO‐mediated intracellular reactive oxygen species accumulation and hindered the generation of the lipid peroxidation product malondialdehyde. Conversely, ferric ammonium citrate notably intensified its cytotoxic effects, especially on retinoic acid (RA)‐resistant SK‐N‐AS cells. Subsequently, the quantitative real‐time polymerase chain reaction results showed ATO significantly inhibited the transcription of GPX4 in NB cells. Remarkably, immunoblotting analysis revealed that MG132 exhibited a notable effect on elevating GPX4 levels in NB cells. Nevertheless, pretreatment with MG132 failed to reverse the ATO‐mediated decrease in GPX4 levels. These findings suggested that ATO reduced the GPX4 expression level in NB cells by mediating GPX4 transcriptional repression rather than facilitating ubiquitinated degradation. In conclusion, our research has successfully indicated that ATO could induce ferroptosis and initiate lipid peroxidation by regulating the transcriptional repression of GPX4, and ATO holds promise as a potential anti‐tumor agent in NB, specifically for patients with RA‐resistant HR‐NB.

Abbreviation

APL

acute promyelocytic leukemia

ATO

arsenic trioxide

DEPC

diethylpyrocarbonate

DFO

deferoxamine

FAC

ferric ammonium citrate

FCM

flow cytometry

FDA

The US Food and Drug Administration

Fer‐1

ferrostatin‐1

GPX

glutathione peroxidase

GSH

glutathione

HR‐NB

high‐risk neuroblastoma

HSCT

hematopoietic stem cell transplantation

MDA

malondialdehyde

MDR

multidrug resistance

MRD

microscopic residual disease

NB

neuroblastoma

Nec‐1

Necrostatin‐1

OS

overall survival

ROH

hydroxyl compounds

ROOH

organic hydroperoxides

ROS

reactive oxygen species

WB

Western blot

Z‐VAD

Z‐VAD (OH)‐FMK

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Neuroblastoma (NB) is the most common extra‐cranial solid tumor in children, originating from primitive neuroblasts and can develop in any region of the sympathetic nervous system. It accounts for ~7%–8% of the incidence of childhood malignancies and contributes to 15% of cancer‐related mortality in children. There is an urgent requirement to develop novel and cost‐effective treatment alternatives that are efficient for neuroblastoma oncogenesis. Arsenic trioxide (ATO) has been the major ingredient of a longstanding traditional Chinese herbal preparation for thousands of years and is now widely adopted in the clinical management of acute promyelocytic leukemia with a favorable safety profile.

• WHAT QUESTION DID THIS STUDY ADDRESS?

This study aims to examine the impact of ATO on the proliferation, migration, and ferroptosis of NB cells. Additionally, it seeks to elucidate the specific mechanism underlying ATO‐induced ferroptosis of NB cells in vitro. Furthermore, this research endeavors to establish a theoretical foundation for the clinical application of ATO in NB treatment.

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

ATO significantly inhibited the proliferation and migration ability of NB cells, and induced ferroptosis. Furthermore, our research has successfully indicated that ATO could induce ferroptosis and initiate lipid peroxidation in NB cells by regulating the transcriptional repression of GPX4. Additionally, the iron compound ferric ammonium citrate could effectively potentiate the cytotoxic effects of ATO on NB cells, particularly on retinoic acid (RA)‐resistant SK‐N‐AS cells.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

ATO, a newly discovered GPX4 inhibitor, holds promise as a potential antitumor agent in NB, specifically for patients with RA‐resistant HR‐NB.

INTRODUCTION

Neuroblastoma (NB) is the most common extra‐cranial solid tumor in children, originating from primitive neuroblasts and can develop in any region of the sympathetic nervous system. It accounts for ~7%–8% of the incidence of childhood malignancies and contributes to 15% of cancer‐related mortality in children. ¹ NB exhibits notable biological and clinical heterogeneity, as low‐risk children have a favorable prognosis, and, in some cases, tumors can even spontaneously regress. ² However, children with high‐risk neuroblastoma (HR‐NB) have a high degree of malignancy, is characterized by a more aggressive course with a higher likelihood of early distant metastasis, accounting for ~40% of all NB cases. ³ Despite a multidisciplinary sequential comprehensive treatment regimen involving chemotherapy, radiotherapy, surgery, immunotherapy, and hematopoietic stem cell transplantation (HSCT), the 5‐year survival rate for patients with HR‐NB remains below 50%. ⁴ Additionally, even if patients initially responded well to intensified chemotherapy, the emergence of multidrug resistance (MDR) and microscopic residual disease (MRD) during later stages of chemotherapy resulted in a 2‐year relapse rate of 80% for HR‐NB. ^{5 , 6} Despite the fact that HSCT and immunotherapy drugs, specifically anti‐GD2 antibodies, have demonstrated substantial enhancements in the 5‐year overall survival (OS) rates among patients with HR‐NB, ⁷ , ⁸ , ⁹ , ¹⁰ the ultimate effectiveness is closely associated with the extent of MRD control in patients after chemotherapy. ¹⁰ , ¹¹ , ¹² Consequently, chemotherapy remains the main method for treating HR‐NB, and cost‐effective and efficient treatment alternatives for children with HR‐NB is urgently needed.

Ferroptosis is a form of programmed cell death that is distinct from apoptosis, and it is characterized by the iron‐dependent peroxidation of lipids. Unlike apoptosis, ferroptosis does not involve the activation of caspase proteases or the requirement of ATP. Instead, it relies on the generation of reactive oxygen species (ROS) induced by iron. ¹³ , ¹⁴ The glutathione peroxidase (GPX) system plays a vital role in mitigating the intracellular buildup of toxic lipid peroxidation and safeguarding against ferroptosis. This system encompasses various essential elements, namely GPX4, glutathione (GSH), and xCT systems. Among them, GPX4 serves as a central antioxidant enzyme that maintains lipid bilayer homeostasis and suppresses the onset of ferroptosis by converting hydroperoxides to water by using reduced GSH as a hydrogen donor and also reducing toxic organic hydroperoxides (ROOHs) to non‐toxic hydroxyl compounds (ROH). ¹⁵ , ¹⁶ Ferroptosis inducer presents a promising approach for eliminating drug‐resistant cancer cells, primarily due to the higher iron levels observed in tumor cells compared to non‐malignant cells. ¹⁷ In recent years, multiple studies have substantiated the potential of ferroptosis inducers as innovative therapeutic strategies for cancer treatment. Particularly noteworthy is the demonstration that sorafenib effectively triggers ferroptosis in hepatocellular carcinoma cells, whereas iron chelators counteract its cytotoxic effects. ¹⁸ Lyuzosulfapyridine has demonstrated the ability to induce ferroptosis in gliomas and head and neck tumors. ^{19 , 20} Furthermore, Dramatoxin A has been found to induce ferroptosis in NB cells by elevating intracellular levels of labile iron and inactivating GPX4. ²¹

Arsenic trioxide (ATO) is the major ingredient of a longstanding traditional Chinese herbal preparation for thousands of years. Clinical studies have demonstrated that low doses of ATO can induce complete remission in patients with relapsed acute promyelocytic leukemia (APL) with favorable safety profile, and are now widely adopted in the clinical management of APL. ²² The utilization of ATO for the treatment of APL was officially sanctioned by the US Food and Drug Administration (FDA) in 2000. ²³ Furthermore, numerous investigations and clinical trials have provided substantial evidence supporting the effectiveness of conventional doses of ATO in eliminating various types of tumor cells, including NB, multiple myeloma, breast cancer, ovarian cancer, hepatocellular carcinoma, and osteosarcoma, etc. ²⁴ , ²⁵ , ²⁶ The antitumor mechanisms of ATO encompass a multitude of aspects, including the induction of differentiation in tumor cells, suppression of telomerase activity, and inhibition of angiogenesis. ²⁷ The findings from our prior investigations have established that the administration of ATO exhibits a notable capacity to augment the cytotoxic effects on NB cells. Moreover, subsequent clinical trials have provided further evidence that the combination of ATO and chemotherapy substantially improves complete response (CR) rates for patients with HR‐NB, whereas concurrently maintaining a favorable safety profile. ²⁸ , ²⁹ Additionally, a previous study by our research team uncovered a substantial enrichment of differential proteins within the ferroptosis pathway subsequent to the administration of ATO to NB cell lines, as evidenced by off‐label quantitative proteomics methodologies. Notably, the expression of GPX4, a crucial rate‐limiting enzyme closely associated with the ferroptosis pathway, was markedly reduced. ³⁰ However, the specific underlying mechanism remains elusive. Herein, we discovered that ATO has the ability to induce ferroptosis and inhibit NB cell proliferation. Subsequent investigations revealed that ATO reduced the GPX4 expression level in NB cells by mediating GPX4 transcriptional repression, rather than by facilitating ubiquitinated degradation. These findings suggest that ATO is a novel GPX4 inhibitor, making it a potential candidate in the development of GPX4 inhibitors for ferroptosis‐based NB therapy.

METHODS

Cell culture

Asbio Technology (Guangzhou, China) provided the human SK‐N‐AS and SH‐SY5Y cells. Cells were cultured in DMEM medium (Gibco, USA) supplemented with 10% heat‐inactivated fetal bovine serum, 100 μg/mL streptomycin, and 100 μg/mL penicillin, in an incubator with 37°C and 5% CO₂. There was no contamination with mycoplasma in any of the cells.

Cytotoxicity assays

The Cell Counting Kit 8 (CCK‐8; K1018, Apexbio, USA) was used to measure the cell growth inhibition. The logarithmic cells were inoculated in 96‐well with (1~2) × 10⁴ per well in 100 μL medium. After 12 h of adherent growth, each cell line was treated with ATO (Harbin Yida Pharmaceutical, Ltd.) at a concentration gradient. NB cells were treated with serial dilutions of ATO for 24 h. Then, cells were transferred into DMEM medium containing CCK‐8 solution at 37°C for 3 h. The enzyme‐linked immunosorbent assay microplate reader (Molecular Devices LLC) was used to access the absorbance values at 450 nm. The half‐maximal inhibitory concentration values of NB cells were calculated using GraphPad Prism version 9.0.

Wound healing assays

Wound healing assays were conducted in order to assess cell migration. NB cells were cultivated in six‐well culture plates until reaching complete confluence. Subsequently, the confluent cell monolayer was gently scraped using a sterile 200 μL pipette tip and exposed to the specified dosage of ATO. The resulting scratches were then observed, and images were captured at 0 and 24 h timepoints, respectively.

Colony formation assays

For colony formation assays, NB cells were cultivated in six‐well plates with 3.0 × 10³ cells/well. After 14‐day incubation, the plates were washed with phosphate‐buffered solution (PBS) twice, fixed by 4% paraformaldehyde fix solution (BL539A; Biosharp, China) for 15 min and stained with 0.1% crystal violet staining solution (C0121; Beyotime, China) for 10 min. Then, we washed off the excess crystal violet solution, dried the six‐well plates, and photographed them.

Lipid peroxidation malondialdehyde assays

The malondialdehyde (MDA) level of NB cells was quantified using an MDA assay kit (S0131S, Beyotime; China). Following treatment with serial dilutions of ATO for 24 h, the cells were lysed using Lysis Solution (P0013; Beyotime, China) and subsequently centrifuged. The MDA working solution was prepared according to the manufacturer's instructions. The test solution was then subjected to boiling for 40 min, cooled to room temperature using running water, and centrifuged to obtain the supernatant. Finally, the absorbance at 532 nm was measured using a multimode reader (Molecular Devices LLC).

ROS assays

Flow cytometry (FCM; BD Biosciences, USA) and an ROS assay kit (Beyotime, China) were used to assess the level of ROS in NB cells. The cells were seeded in six‐well plates with 3 mL of DMEM medium at a density of 8 × 10⁵ cells overnight and treated with various dilutions of ATO for 12 h. Subsequently, the cells were transferred to serum‐free medium containing the fluorescent probe DCFH‐DA (2000:1) and incubated in the dark at 37°C for 20 min. Afterward, the cells were washed three times with PBS. The green fluorescence intensity was quantified using an automated fluorescent microplate reader (Scientific Varioskan LUX; Thermo Fisher Scientific).

Measurements of intracellular iron

Flow cytometry (FCM; BD Biosciences, USA) and an iron assay kit (F374; Dojindo, China) were used to assess the intracellular iron level in NB cells. The cells were seeded in six‐well plates with 3 mL of DMEM medium at a density of 8 × 10⁵ cells overnight and treated with various dilutions of ATO for 12 h. Subsequently, the cells were transferred to serum‐free medium containing the fluorescent probe FerroOrange (1000:1) and incubated in the dark at 37°C for 30 min. Afterward, the cells were washed three times with PBS. The fluorescence intensity was quantified using an automated fluorescent microplate reader (Scientific Varioskan LUX; Thermo Fisher Scientific).

Ferroptosis rescue experiment analysis

Ferroptosis rescue experiment analysis was performed by CCK‐8. Following an overnight culture of NB cells, they were subjected to pretreatment with specific compounds including deferoxamine (DFO; 100 μM), rerrostatin‐1 (Fer‐1; 1 μM), Z‐VAD (OH)‐FMK (Z‐VAD, 10 μM), and Necrostatin‐1 (Nec‐1, 10 μM) for 1 h. Subsequently, the cells were were treated with ATO at a concentration of 20 μM for 24 h. Subsequently, the cells were evaluated in accordance with the manufacturer's instructions. The absorbance values at 450 nm were measured using an enzyme‐linked immunosorbent assay microplate reader.

Quantitative real‐time polymerase chain reaction analysis

The total RNA extraction of NB cells was conducted using Vinozane reagent in conjunction with chloroform, following the specified protocol subsequent to a 24‐h drug treatment. We prepared RNA precipitates using isopropanol, washed them with ethanol to a concentration of 75%, and then dissolved in water treated with diethylpyrocarbonate. Total RNA purity and integrity was examined by measuring the OD₂₆₀ and OD₂₈₀ of the extracted RNA and their ratios using an enzyme marke. The cDNA was synthesized using HiScipt II One Step RT‐PCR Kit (Vazyme Biotech, China). Amplification of the cDNA was performed following cDNA synthesis to determine the number of transcripts for GPX4. In order to perform the measurements, a Biometra T‐gradient Thermoblock thermal cycler (Gottingen, Germany) was used. The real‐time polymerase chain reaction (RT‐PCR) multiplex was performed using a total volume of 20 μL, with each set of forward and reverse primers being applied to a volume of 1, 10 μL of ChamQ universal qPCR premix (Vazyme Biotech, China) being added, 7 μL of pure water, and 1 μL of cDNA being added. In the initial cycle of the PCR, the polymerase was activated at 95°C for 10 min, then 40 cycles of PCR amplification were performed, with the denaturation step conducted at 95°C for 10 s, the annealing step conducted at 60°C for 15 s, and the elongation step taken place at 72°C for 20 s. GAPDH was used to normalize mRNA expression. Each individual PCR experiment was conducted using the same conditions as those used during the multiplex experiments. The multiplex PCR primer sequences are listed in Table 1.

TABLE 1.

PCR primer sequences.

Gene	Forward primer	Reverse primer
GPX4	5′‐GAGGCAAGACCGAAGTAAACTAC‐3′	5′‐CCGAACTGGTTACACGGGAA‐3′
GAPDH	5′‐GGTGGTCTCCTCTGACTTCAACA‐3′	5′‐GTTGCTGTAGCCAAATTCGTTGT‐3′

Abbreviation: PCR, polymerase chain reaction.

Western blot analysis

The GPX4 protein level of NB cells was measured using Western blot (WB) analysis. The cells were seeded in six‐well plates with 3 mL of DMEM medium at a density of 1 × 10⁶ cells overnight. Following pretreatment with the proteasome inhibitor MG‐132 for 1 h, NB cells in each experimental group were treated with various concentrations of ATO for 24 h. Protein lysates were prepared using lysis buffer at 4°C, and the protein concentration was determined using the BCA protein assay (Beyotime, China). Following SDS‐PAGE separation, the samples were transferred onto PVDF membranes and washed with TBST and blocked with TBST containing 5% milk powder. The primary antibodies (Cell Signaling Technology) targeting GPX4 were diluted in a 5% BSA/TBST solution and incubated overnight at 4°C. A horseradish peroxidase‐labeled antibody (Sigma Chemical) was used as the secondary antibody to label the membranes for 1 h at 37°C. Subsequently, the membranes were subjected to three washes with TBST for a duration of 10 min each. Image J software was used to quantify the protein levels after nitrocellulose membranes were coated with proteins.

Statistical analysis

The data was analyzed by GraphPad Prism version 9.0 (GraphPad Software). Statistical data are presented as mean ± standard deviation (SD). Normality and variance homogeneity were assessed using Levene test and Kolmogorov–Smirnov test separately. A Student's t‐test was used when the quantitative data satisfied the above conditions, whereas a rank‐sum test was used when they did not. We considered statistical significance at p < 0.05.

RESULTS

ATO inhibits the viability and migration of NB cells

To explore the effect of ATO on the killing of NB cell lines, we examined the cytotoxicity of ATO on NB cells SK‐N‐AS and SH‐SY5Y. The CCK‐8 assays revealed that ATO inhibited the proliferation of SK‐N‐AS and SH‐SY5Y cells in a dose‐dependent manner, and ATO was more cytotoxic to SH‐SY5Y cells than to retinoic acid (RA)‐resistant strain SK‐N‐AS cells (Figure 1a,b). To verify the effect of ATO on the cell proliferation and migration of SK‐N‐AS and SH‐SY5Y cells, we performed colony‐forming assays and wound healing assays, which showed that ATO inhibits the proliferation and migration ability of NB cell lines (Figure 1c–g). Taken together, ATO significantly inhibited the proliferation and migratory ability of NB cell lines and remained highly cytotoxicity against RA‐resistant cells.

FIGURE 1.

The cytotoxic effect of ATO on NB cells. (a, b) The cell viability of NB cells treated by ATO were examined by CCK‐8 assay. (c–g) Colony‐forming assays and wound healing assays for the ATO and control groups in SH‐SY5Y and SK‐N‐AS cells indicated that ATO inhibited the proliferation and migration ability. Data are presented as the mean ± SD, n = 3, ****p < 0.0001, compared with the control group (0 μM). ATO, arsenic trioxide; NB, neuroblastoma.

ATO induces ferroptosis in NB cells

For examining whether ATO induced ferroptosis in NB cell lines, we examined the intracellular levels of lipid ROS after ATO treatment of SK‐N‐AS and SH‐SY5Y cells. FCM experiments showed that ATO treatment of both NB cell lines for 12 h significantly increased intracellular ROS levels (Figure 2a–d). Further experiments confirmed that ATO promotes the accumulation of MDA (Figure 2e,f), which is a frequently investigated oxidative stress marker for lipid peroxidation assessment. Additionally, our investigation unveiled a noteworthy augmentation in intracellular iron levels in both NB cell lines following ATO treatment (Figure 2g,i), providing further theoretical support for the induction of ferroptosis in NB cells by ATO.Together, our above experiments confirmed that ATO induces ferroptosis in NB cells.

FIGURE 2.

ATO facilitated the lipid peroxidation and induced the ferroptosis in NB cells. (a–d) NB cells were treated with different concentrations of ATO for 12 h, then the lipid ROS (DCFH‐DA 581/591 fluorescence staining) were assessed by FCM. (e, f) NB cells were treated with different concentrations of ATO for 24 h, the levels of MDA were measured by a MDA assay kit. (g–i) NB cells were treated with different concentrations of ATO for 12 h, then the intracellular iron levels were assessed by FCM. Data are presented as the mean ± SD, n = 3, ****p < 0.0001, compared with control group (0 μM). ATO, arsenic trioxide; FCM, flow cytometry; MDA, malondialdehyde; NB, neuroblastoma; ROS, reactive oxygen species.

Part of the ATO‐mediated cytotoxicity is dependent on ferroptosis

To examine the impact of ATO on ferroptosis, we assessed the cytotoxicity of ATO on NB cells while utilizing the iron chelator DFO, the ferroptosis inhibitor Fer‐1, the apoptosis inhibitor Z‐VAD, and the necrosis inhibitor Nec‐1. CCK‐8 results (Figure 3a,b) suggested that ATO combined with the iron chelator DFO partially rescues ATO‐induced NB ferroptosis, whereas ATO combined with Nec‐1 or Z‐VAD does not rescue ATO‐induced NB cell death. Interestingly, ATO combined with Fer‐1, an ferroptosis inhibitor, also failed to rescue ATO‐induced ferroptosis. Further experiments confirmed that DFO reversed ATO‐mediated intracellular ROS accumulation and inhibited ATO‐induced generation of lipid peroxidation product MDA (Figure 3c–h). Furthermore, in order to delve deeper into the impact of ATO on ferroptosis, we conducted an evaluation of the viability of NB cells in the presence of ferric ammonium citrate (FAC). Intriguingly, when compared to the cytotoxic effect of ATO alone, the concurrent administration of FAC notably intensified the cytotoxic effect of ATO in NB cells (Figure 3i–l). Therefore, the results of the foregoing studies show that ATO is indeed capable of inducing ferroptosis in the NB cells. SK‐N‐AS and SH‐SY5Y and that DFO inhibited ATO‐mediated cytotoxicity, whereas FAC augmented ATO‐mediated cytotoxicity.

FIGURE 3.

ATO‐mediated cytotoxicity is dependent on ferroptosis. (a, b) NB cells were incubated with different concentrations of ATO in the presence of Fer‐1 (1 μM) or DFO (100 μM) or Z‐VAD (10 μM) or Nec‐1 (10 μM), then cell viability of NB cells were measured by CCK‐8 assay. (c–f) NB cells were cotreatment with DFO for 12 h, then the lipid ROS (DCFH‐DA 581/591 fluorescence staining) were assessed by FCM. (g, h) NB cells were cotreatment with DFO for 24 h, then the levels of MDA were measured by a MDA assay kit. (i–l) The cell viability of NB cells co‐treated with FAC or FAC alone was examined by a CCK‐8 assay, and the concurrent administration of (i) FAC (100 μM) or (k) FAC (200 μM) notably intensified the cytotoxic effect of ATO in NB cells. Data are presented as the mean ± SD, n = 3, **p < 0.01; ***p < 0.001; ****p < 0.0001; ns: no significance. ATO, arsenic trioxide; DFO, deferoxamine; FCM, flow cytometry; IC₅₀, half‐maximal inhibitory concentration; MDA, malondialdehyde; NB, neuroblastoma; ROS, reactive oxygen species.

ATO‐mediated ferroptosis is dependent on GPX4 transcriptional repression

GPX4, a widely recognized antioxidant enzyme within the GPX system, plays a crucial role in impeding ferroptosis, which maintains the homeostasis of the cellular lipid bilayerby reducing toxic ROOH to non‐toxic ROH by using reduced GSH as a hydrogen donor. ¹⁴ Our previous study confirmed that the intracellular level of GPX4 protein significantly decreased after ATO treatment in NB cell lines, as revealed by off‐label quantitative proteomics techniques, however, the specific mechanism is still unclear. ³⁰ As shown in Figure 4a,b, quantitative RT‐PCR (qRT‐PCR) results showed that ATO significantly inhibited GPX4 transcription in NB cells. To detect whether ATO promotes the degradation of GPX4, we investigated the effect of ATO on GPX4 based on NB cells pretreated by a proteasome inhibitor MG132. WB results showed that MG132 increased GPX4 levels in NB cells, suggesting that GPX4 is degraded by the ubiquitinated proteasome pathway. Simultaneously, MG132 did not reverse the ATO‐mediated decrease in GPX4 intracellular levels (Figure 4c–f). Overall, these findings demonstrated that ATO can mediate GPX4 transcriptional repression in NB cell lines.

FIGURE 4.

ATO promoted the transcriptional repression of GPX4 in NB cells. (a, b) The expression of GPX4 in NB cells were measured by qRT‐PCR after treatment with different concentrations of ATO for 24 h. (c–f) NB cells were incubated with MG132 (10 μM) for 2 h, then with or without different concentrations of ATO for 24 h, and the expression of GPX4 was measured by immunoblotting. Data are presented as the mean ± SD, n = 3, *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns: no significance, compared with control group (0 μM). ATO, arsenic trioxide; NB, neuroblastoma; qRT‐PCR, quantitative real‐time polymerase chain reaction.

DISCUSSION

NB is the most common extracranial solid tumor in children, contributing to ~15% of childhood cancer‐related mortality. ¹ A significant proportion, nearly 40%, of NB cases are classified as HR‐NB. Despite the utilization of a comprehensive treatment approach involving chemotherapy, radiotherapy, surgery, immunotherapy, and HSCT, the 5‐year survival rate for HR‐NB remains below 50%. ⁴ The emergence of MDR and MRD in the late stage of chemotherapy decreases the efficacy of HSCT and immunotherapy resulting in the 2‐year relapse rate as high as 80%. ⁵ Therefore, more effective chemotherapeutic agents, to improve the treatment outcomes and survival rates for patients with HR‐NB, are urgent and imperative.

ATO has been used medicinally for therapeutic purposes for more than 2 millennia in traditional Chinese medicine, exhibiting noteworthy clinical implications in the realm of oncology treatment. A growing corpus of evidence indicates that ATO possesses the capacity to initiate apoptosis in neoplastic cells via various mechanisms, such as augmenting intracellular levels of ROS, thereby leading to irreversible impairment of the mitochondrial membrane potential, ³¹ activating CASPASE3/7 to initiate apoptosis, ³² facilitating tumor cell differentiation, and impeding the cell cycle progression, ³³ , ³⁴ and so on. Previous research has demonstrated that ATO effectively arrests the progression of NB cells in the G2/M phase, while concurrently augmenting the cytotoxicity of M phase‐specific chemotherapy drugs on SK‐N‐SH cells. ³³ Moreover, ATO in combination with DNMTi analogs like decitabine could enhance cytotoxic effects on SK‐N‐SH cells, and continued administration did not induce increased expression of the drug resistance protein P‐gp. ³⁵ However, a phase II clinical study conducted by Shakeel Modak et al. revealed that the combination of ATO and (131)I‐MIBG for the treatment of relapsed or refractory stage 4 NB or metastatic pheochromocytoma did not yield a higher response rate compared to the use of single‐agent (131)I‐MIBG, despite the patients' favorable tolerance of the therapy. ³⁶ In contrast, our multicenter non‐randomized controlled clinical study has provided evidence that the combination of ATO and chemotherapy, in comparison to conventional chemotherapy, could significantly enhance the objective remission rate (89.74% vs. 46.15%) and CR (69.23% vs. 23.08%) in patients with HR‐NB, while also being well‐tolerated. ²⁹

In this study, we initially discovered ATO, a natural small molecule, as a novel GPX4 inhibitor showed a notable cytotoxic effect on NB cells, including SH‐SY5Y and RA‐resistant cell lines SK‐N‐AS, which may potentially serve as a promising therapeutic approach for patients with HR‐NB. Our study confirmed that ATO significantly inhibited the proliferation and migration ability of NB cells, and induced ferroptosis. In addition, the iron chelator DFO partially reversed the cytotoxic effect of ATO on NB cells, whereas FAC significantly augmented its cytotoxic effects, especially on RA‐resistant SK‐N‐AS cells, suggesting that the ferroptosis might be the main mechanism inducing cytotoxicity in the NB cells exposured to ATO. Interestingly, ATO combined with the ferroptosis inhibitor Fer‐1 did not rescue ATO‐induced ferroptosis, whereas DFO reversed ATO‐mediated intracellular ROS accumulation and inhibited the production of lipid peroxidation product MDA. Notably, these observations may be related to the mechanism by which ATO‐induced ferroptosis.

GPX4, a central regulator of intracellular resistance to lipid peroxidation processes, decreased intracellular levels of which can lead to elevated intracellular lipid ROS and ferroptosis, thereby inhibiting cell proliferation. ³⁷ It was found that GPX4 inhibitors RSL3, ³⁸ baicalin, ³⁹ and baileyanin ⁴⁰ could induce ferroptosis in tumor cells by inhibiting GPX4. To further clarify the specific mechanism by which ATO mediates the reduction of intracellular GPX4 protein levels, our qRT‐PCR results showed that ATO significantly inhibited the transcription of GPX4 in NB cell lines SK‐N‐AS and SH‐SY5Y, indicating that ATO inhibited GPX4 synthesis. There had been reported that the proteasome inhibitor MG132 could lead to intracellular GPX4 accumulation, suggesting that GPX4 may regulate degradation through the ubiquitinated proteasome pathway. ⁴¹ Notably, immunoblotting analysis revealed that MG132 exhibited a notable effect on elevating GPX4 levels in NB cells. Nevertheless, the pretreatment of MG132 was ineffective in reversing the decline of GPX4 levels induced by ATO in NB cells. Furthermore, no significant disparity was observed when ATO was administered alone or in conjunction with MG132. These findings suggested that ATO reduced the GPX4 expression level in NB cells by mediating GPX4 transcriptional repression rather than facilitating ubiquitinated degradation. This may also explain why the iron chelator DFO partially reversed the cytotoxic effects of ATO on NB cells, but the ferroptosis inhibitor Fer‐1 did not rescue ATO‐induced ferroptosis. Giovanni Miotto et al. demonstrated that Fer‐1 exerted an inhibitory effect on ferroptosis by scavenging the initial intracellular alkoxy radicals and elevating GSH levels. ⁴² , ⁴³ The process of intracellular GSH against ferroptosis required the involvement of the rate‐limiting enzyme GPX4. In the ATO‐treated NB cells, the transcription of GPX4 was markedly suppressed, leading to a notable reduction in the protein expression level of GPX4. Therefore, despite Fer‐1's ability to scavenge alkoxy radicals and elevate GSH levels, the absence of the rate‐limiting enzyme GPX4 in NB cells hindered GSH from exerting its anti‐lipid peroxidation properties, ultimately rendering Fer‐1 ineffective in inhibiting the cytotoxic impact of ATO on NB cell lines.

In conclusion, our study has successfully shown that ATO, a newly discovered GPX4 inhibitor, has the ability to induce ferroptosis in NB cells by regulating the transcriptional repression of GPX4 (Figure 5). Furthermore, the iron compound FAC could effectively potentiate the cytotoxic effects of ATO on NB cells, particularly on RA‐resistant SK‐N‐AS cells. This research has laid the foundation for further advancing the clinical application of ATO in HR‐NB, specifically for patients with RA‐resistant HR‐NB. It presents a promising novel therapeutic strategy that can potentially enhance the efficacy of NB treatment and improve the OS rate among children with HR‐NB.

FIGURE 5.

Schematic diagram of the proposed mechanism by which ATO induce ferroptosis by decreasing intracellular GPX4 protein expression levels. ATO induces cellular ferroptosis by mediating GPX4 transcriptional repression in NB cells, leading to intracellular accumulation of lipid peroxidation products. ATO, arsenic trioxide.

AUTHOR CONTRIBUTIONS

M.S., and X.L. wrote the manuscript. M.S., and Y.L. designed the research. M.S., X.L., Y.M., X.P., and X.X. performed the research. M.S., and X.L. analyzed the data. W.W., K.H., and Y.L. contributed new reagents/analytical tools.

FUNDING INFORMATION

This work was supported by the Guangzhou Area Clinical Specialty Technology Program (Grant Number 2023P‐TS39), Sun Yat‐Sen Clinical Research Cultivating Program (Grant Number SYS‐C‐202007), Sun Yat‐Sen Medical‐industrial Integration Cultivating Program (Grant Number YXYGRH202203), Heilongjiang Harbin Medical University Pharmaceutical Co., Ltd. (Grant Number 7670020013) and Traditional Chinese Medicine Bureau of Guangdong Province (Grant Number 20221078).

CONFLICT OF INTEREST STATEMENT

The authors declared no competing interests for this work.

Su M, Liu X, Ma Y, et al. Arsenic trioxide induces ferroptosis in neuroblastoma by mediating GPX4 transcriptional inhibition. Clin Transl Sci. 2024;17:e13716. doi: 10.1111/cts.13716