ARTICLE OPEN

Cellular and Molecular Biology

GSTM3 enhances radiosensitivity of nasopharyngeal carcinoma

by promoting radiation-induced ferroptosis through USP14/

FASN axis and GPX4

Yuting Chen 1,4, Yuanyuan Feng1,4, Yanling Lin1,4, Xiaohan Zhou1

, Lingzhi Wang2

, Yingtong Zhou1

, Kefan Lin3 and Longmei Cai 1✉

© The Author(s) 2024

BACKGROUND: Radiotherapy is a critical treatment modality for nasopharyngeal carcinoma (NPC). However, the mechanisms

underlying radiation resistance and tumour recurrence in NPC remain incompletely understood.

METHODS: Oxidised lipids were assessed through targeted metabolomics. Ferroptosis levels were evaluated using cell viability,

clonogenic survival, lipid peroxidation, and transmission electron microscopy. We investigated the biological functions of

glutathione S-transferase mu 3 (GSTM3) in cell lines and xenograft tumours. Co-immunoprecipitation, mass spectrometry, and

immunofluorescence were conducted to explore the molecular mechanisms involving GSTM3. Immunohistochemistry was

performed to investigate the clinical characteristics of GSTM3.

RESULTS: Ionising radiation (IR) promoted lipid peroxidation and induced ferroptosis in NPC cells. GSTM3 was upregulated

following IR exposure and correlated with IR-induced ferroptosis, enhancing NPC radiosensitivity in vitro and in vivo.

Mechanistically, GSTM3 stabilised ubiquitin-specific peptidase 14 (USP14), thereby inhibiting the ubiquitination and subsequent

degradation of fatty acid synthase (FASN). Additionally, GSTM3 interacted with glutathione peroxidase 4 (GPX4) and suppressed

GPX4 expression. Combining IR treatment with ferroptosis inducers synergistically improved NPC radiosensitivity and suppressed

tumour growth. Notably, a decrease in GSTM3 abundance predicted tumour relapse and poor prognosis.

CONCLUSIONS: Our findings elucidate the pivotal role of GSTM3 in IR-induced ferroptosis, offering strategies for the treatment of

radiation-resistant or recurrent NPC.

British Journal of Cancer; https://doi.org/10.1038/s41416-024-02574-1

BACKGROUND

Nasopharyngeal carcinoma (NPC) is a malignancy originating from

the nasopharyngeal epithelium and commonly diagnosed in

southern China, Southeast Asia, and North Africa [1]. Radiotherapy

plays a crucial role in the treatment of patients with nonmetastatic NPC [2]. While the 5-year overall survival rate for NPC

patients who have undergone standardised treatment has

increased to 80–90%, approximately 10% of patients experience

local recurrence due to radiotherapy resistance [3–5]. Radioresistance is attributed to diverse epigenetic regulatory mechanisms, including epithelial–mesenchymal transition, cancer stem

cell properties, autophagy, and the oncogenic metabolism

microenvironment [6–9]. However, the current understanding of

these mechanisms does not fully address the challenges posed by

radioresistance and tumour recurrence in NPC. Therefore, it is

imperative to explore novel mechanisms that contribute to

radioresistance and identify promising strategies to enhance the

treatment response of NPC patients.

Ferroptosis is a distinct form of programmed cell death

characterised by the involvement of intracellular iron and

excessive lipid peroxidation, morphologically and mechanistically

separating from other forms of cell death [10, 11]. Notably,

ferroptosis plays a critical role in modulating radiotherapy

sensitivity through diverse regulatory mechanisms in various

cancers, including hepatocellular carcinoma, lung cancer, and

melanoma [12–14]. Several classes of ferroptosis inducers (FINs)

have been extensively studied and shown to effectively modulate

cancer progression, regulate tumour microenvironment, and

enhance treatment response [15–17]. Class I FINs, such as erastin

and sorafenib, exert their effects by suppressing solute carrier

family 7 member 11 (SLC7A11), limiting intracellular cysteine

uptake and subsequent glutathione synthesis to induce ferroptosis. Class II and Class III FINs impede glutathione peroxidase 4

(GPX4) enzymatic activity and deplete GPX4 protein, respectively,

hindering glutathione conversion and lipid hydroperoxide elimination. Targeting GPX4- or SLC7A11-induced ferroptosis regulates

Received: 11 July 2023 Revised: 28 December 2023 Accepted: 3 January 2024

1

Department of Radiation Oncology, Nanfang Hospital, Southern Medical University, 510515 Guangzhou, China. 2

Department of General Surgery, Nanfang Hospital, Southern

Medical University, 510515 Guangzhou, China. 3

First Clinical Medical College, Southern Medical University, 510515 Guangzhou, China. 4

These authors contributed equally: Yuting

Chen, Yuanyuan Feng, Yanling Lin. ✉email: clm520@i.smu.edu.cn

British Journal of Cancer www.nature.com/bjc

1234567890();,:

tumour development and enhances treatment response, offering

a promising therapeutic approach for cancers [18, 19]. Itraconazole, cephalosporin, disulfiram/copper, and cucurbitacin B have

been identified to possess potential antitumour activity in NPC by

triggering ferroptosis [20–23]. However, the role of ferroptosis in

sensitising NPC to radiotherapy has not been investigated in

previous studies.

Glutathione S-transferase mu 3 (GSTM3), a member of the

glutathione S-transferase family, exerts diverse effects on the

progression of various malignancies [24]. On the one hand, GSTM3

influences the malignant metabolic pattern in pancreatic cancer,

alleviates aggressiveness in renal cell carcinoma, and reverses

radioresistance in hepatocellular carcinoma to suppress tumour

malignancy [25–27]. On the other hand, specific genotypes of

GSTM3 may confer increased susceptibility to cervical, colorectal,

and prostate cancers [28–30]. Additionally, GSTM3 is associated

with the malignant tumour behaviours and poor prognosis in

colon cancer and glioma [31, 32]. However, the regulatory effects

and detailed mechanisms of GSTM3 in NPC remain unclear.

In this study, we uncovered the crucial role of GSTM3 in

facilitating ionising radiation (IR)-induced ferroptosis to enhance

radiosensitivity in NPC. Mechanistically, GSTM3 acted by stabilising

the ubiquitin-specific peptidase 14 (USP14)/fatty acid synthase

(FASN) axis and directly inhibiting the expression of the

glutathione peroxidase GPX4. The combination of FINs and IR

treatment synergistically enhanced NPC radiosensitivity and

inhibited tumour growth. Importantly, GSTM3 was correlated with

radiotherapy response and predicted a favourable prognosis in

NPC patients. These findings emphasise the significant role of

GSTM3 in IR-induced ferroptosis and radiotherapy sensitivity,

providing valuable insights for the development of promising

treatment strategies targeting radiation-resistant or recurrent NPC.

METHODS

Cell culture, cell transfection, and establishment of cell lines

Human NPC cell lines (5-8F, HONE1, and CNE2) were generously provided

by Sun Yat-sen University Cancer Center, Guangzhou, China. The cells were

cultured in RPMI-1640 medium (Gibco) supplemented with 10% fetal

bovine serum (Gibco). All cell lines were maintained in a humidified

chamber at 37 °C with 5% CO2.

For transfection experiments, small interfering RNA (siRNA) oligonucleotides (GemmaPharma) or plasmid DNAs (GeneChem) were transfected into

the cells using Lipofectamine 3000 transfection reagent (Invitrogen)

according to the manufacturer’s instructions. The specific sequences of

the siRNA oligonucleotides are shown in Supplementary Table 1.

To establish a radiation-resistant cell line, CNE2 cells were repeatedly

exposed to IR (6 Gy) until the fifth-generation surviving cells were obtained and

designated as CNE2-R. To generate stable cell lines overexpressing GSTM3,

cells were transfected with the Flag-GSTM3-GFP lentivirus synthesised by

GeneChem. Puromycin (2 μg/mL) was used to select for cell colonies

successfully transduced with the GSTM3-overexpressing lentivirus.

Cell viability assays

NPC cells were treated with the designated concentrations of reagents or

exposed to indicated doses of IR using a 6-MV X-ray beam. After 24 h, the

culture medium was replaced with 100 μL fresh medium containing 10%

Cell Counting Kit-8 reagent (FDbio Science), and cells were incubated for

1–4 h in an incubation chamber. The absorbance at 450 nm was measured

using a microplate reader (Bio-Rad), and cell viability was calculated

following the manufacturer’s instructions.

Clonogenic survival assays

NPC cells were seeded in triplicate into six-well plates and exposed to

varying doses of IR. The cells were incubated for 7–10 days until visible

colonies formed. After washing with phosphate-buffered saline (PBS), the

cell colonies were fixed with 4% paraformaldehyde for 15 min. Subsequently, the colonies were stained with 0.5% crystal violet (Macklin) for

30 min. Colonies with more than 50 cells were counted to calculate the

surviving fraction. The curve of survival fraction (SF) with increasing dose

(D) was plotted by the multi-target single-hit model: SF = 1– (1– e[−kD]

) N

.

The radiobiological parameters were derived: D0 = 1/k, Dq = D0 × ln (N),

and SF2 = 1 – (1 – e[−2k]

) N

.

RNA extraction and quantitative reverse transcription PCR

(qRT-PCR)

Total RNA was extracted from cell lysates using TRIzol reagent (TaKaRa Bio)

and reverse-transcribed into cDNA using PrimeScript RT reagent Kit

(TaKaRa Bio). qRT-PCR was performed using SYBR Premix ExTaqSYBR Green

PCR kit (TaKaRa Bio). The Ct values were obtained and analysed using a

LightCycler® 480 (Roche, Basel, Switzerland). β-Actin was used as a loading

control. The primer sequences are listed in Supplementary Table 2.

Lipid peroxidation assay

Cells were treated with FINs (5 μM FIN56 or sorafenib) or subjected to

irradiation. After 24 h, the cells were washed twice with PBS and incubated

with 500 μL of PBS containing 5 μM C11-BODIPY 581/591 dye (Invitrogen)

for 30 min in the dark. The uncombined C11-BODIPY dye was removed by

washing the cells with PBS. The fluorescence emitted by C11-BODIPY 581/

591 was detected by simultaneously measuring the green (484/510 nm)

and red (581/610 nm) signals using a flow cytometer (BD FACSAria III).

Transmission electron microscopy (TEM)

The samples of cells were gently scraped using cell scrapers with 1 mL PBS

and carefully transferred to 1.5 mL tubes. After centrifugation, the cell

pellets were fixed with 500 μL of 2.5% glutaraldehyde at room

temperature. After dehydration, embedding, and preparation of ultrathin

sections, the samples were observed using a Hitachi H-7500 TE

microscope.

High-throughput targeted metabolomics of oxylipins

The cells in the test group were exposed to 6 Gy IR and then collected for

high-throughput targeted metabolomics of oxylipins after 24 h. In brief,

cell samples were processed using an extraction solution containing an

isotopically labelled mixture for metabolite extraction. After homogenisation, sonication, and purification, the purified samples were evaporated to

dryness and dissolved in 30% acetonitrile. The clear supernatant was

subjected to ultrahigh-performance liquid chromatography-tandem mass

spectrometry analysis. SCIEX Analyst Work Station (version 1.6.3) and

Multiquant 3.03 software were employed for data acquisition and

processing. High-throughput targeted metabolomics of oxylipins was

conducted by Biotree Biological Technology (Shanghai, China).

Whole transcriptome sequencing

After mRNA extraction, purification, and fragmentation, mRNAs were

reverse-transcribed into single-stranded cDNA, and then the second strand

of cDNA was synthesised. The cDNA was purified using Ampure Beads XP

(Beckman). Subsequently, the ends of the cDNAs were repaired, and

polyadenylation was added to the 3’-end. Sequencing adaptors were

ligated to the cDNA ends, and PCR amplification was performed. After

quality control, the library was subjected to paired-end sequencing using

the Hiseq 2000 system (Illumina) with technical support provided by

Shanghai Biotechnology Corporation.

Protein extraction and western blot analysis

The pre-processed cells were washed twice with PBS and lysed using

radioimmunoprecipitation assay lysis buffer on ice for 15 min. After

ultrasonic concussion, the cell lysates were separated by centrifugation at

15,000 × g at 4 °C for 15 min. Total protein concentration was quantified

using the bicinchoninic acid assay (Fdbio Science).

For western blot analysis, proteins were separated by sodium dodecyl

sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membranes (Merck Millipore). After blocking with 5%

bovine serum albumin (BSA) for 1 h, the membranes were incubated

overnight at 4 °C with specific primary antibodies, as indicated in

Supplementary Table 3.

Co-immunoprecipitation (Co-IP) and mass spectrometry

analysis

The cells were transfected with plasmids with Flag label. Afterward, cells

were lysed with IP cell lysis buffer (Beyotime) at 4 °C for 30 min. Cell

Y. Chen et al.

2

British Journal of Cancer

lysates were then incubated with anti-Flag antibody (Sigma-Aldrich)

overnight at 4 °C. Protein A/G agarose beads (Santa Cruz Biotechnology)

were added to the samples and incubated at 4 °C for 4 h. The immune

complexes were subsequently washed three times with PBS and eluted

with 2× SDS loading buffer (Invitrogen). The eluates were subjected to

western blot analysis using the indicated antibodies. Liquid

chromatography-mass spectrometry (LC-MS) analysis was performed by

Wininnovate Bio (Shenzhen, China).

5-8F

0.0

1.0

2.0

3.0

4.0

0

0

0.001

10-Nitrooleic acid

±12(13)-EpOME

±9(10)-EpOME

±14(15)-EET

±9(10)-DiHOME

15-OxoETE

±8-HDoHE

+12(13)-DiHOME

+14,(15)-DiHETrE

13-OxoODE

15-OxoEDE

16S-HETE

13S-HOTrE()

DTA

0.01

0.1

1

2468

2 4

ns

6 8

Dose (Gy)

0 2468

Dose (Gy)

0 2 4 6 8

Dose (Gy)

IR

IR+Ferr-1

IR

IR+Ferr-1

Gy

e

g

Ferr-1 CON

GyFerr-1 CON

Surviving fraction

0.001

0.01

0.1

1.0

0.5

0.0

Cell viability

1

Surviving fraction

0.0001

0.01

0.001

0.1

1

Surviving fraction

Relative PTGS2 expression Lipid oxidation level (% of control)

5-8F

HONE1 5-8F

HONE1

HONE1

CNE2

CNE2 CNE2

5-8F

5-8F

CON

d

–1.5

–1

–0.5

0

0.5

1.5

a

1

IR: 6 Gy, 24h

HONE1

HONE1

CNE2

CNE2

0 2 4 6 8 GyFerr-1 CON

0 2 4 6 8

IR+Ferr-1

IR

CON

IR

CON

IR+Ferr-1

IR

CON

0

50

100

150

250

200

IR

IR+Ferr-1

IR

CON

500 nm

±11(12)-EET

±8,9-DiHETrE

±11,12-DiHETrE

±19,20-EpDPE

±11-HDoHE

± 18-HETE

± 20-HDoHE

± 13-HDoHE

± 11-HEPE

5S-HETrE

11S-HETE

13S-HODE

13S-HOTrE

5S-HEPE

9S-HODE

9S-HOTrE

DHA

8S-HETrE

9R-HETE

ARA

15S-HETrE

PGE1

EPA

TXB2

6-trans LTB4

12S-HEPE

9-OxoODE

12S-HETE

5-OxoETE

1a,1b-dihomo PGE2

15S-HEPE

12S-HHTrE

8S,15S-DiHETE

15-deoxy-Δ 12,14-PGJ2

6-keto PGF1 Tetranor-12S-HETE

-log[10](PValue)

Log[2](FC)

b

c

0.0

0.5

1.0

1.5

2.0

f

–1

Sig

Down

Up

15-OxoEDE

15-OxoETE

13S-HOTrE()

11S-HETE ±9(10)-DiHOME

13S-HODE

5S-HEPE

13-OxoODE

16S-HETE

±8-HDoHE

15S-HEPE

13S-HOTrE 9S-HOTrE

±11-HEPE 9S-HODE

±11,12 –DiHETrE 8S-HETrE ±14,15-DiHETrE

DHA 9R-HETE

±12(13)-DiHOME

ARA

tetranor-12S-HETE

±12(13)-EpOME

DTA

6-keto PGF1

1a,1b-dihomo PGE2

10-Nitrooleic Acid

5S-HETrE ±13-HDoHE

15S-HETrE

±9(10)-EpOME 12S-HETE

12S-HEPE

9-OxoODE

±11-HDoHE

±20-HDoHE

5-OxoETE

±19,20-EpDPE EPA

8S,15S-DiHETE

±14(15)-EET

±18-HETE

6-trans LTB4

PGE1

±8,9–DiHETrE

TXB2

15-deoxy-12,14-PGJ2

12S-HHTrE

±11(12)-EET

0 1

Y. Chen et al.

3

British Journal of Cancer

In vitro ubiquitination assay

Cells were transfected with siRNA alongside HA-ubiquitin plasmid. After

48 h of transfection, 20 μM MG-132 (MedChemExpress) was added to

inhibit proteasomal degradation. Immune complexes were obtained by

immunoprecipitation with anti-HA antibody (Dia-An Biotech). The ubiquitination of the designated substrate was then detected via western blot

analysis using the indicated antibodies.

Immunofluorescence (IF)

NPC cells were seeded on slides overnight, then fixed with 4%

paraformaldehyde and permeabilised with 0.5% Triton X-100. Subsequently, the cells were blocked with 5% BSA and incubated with indicated

primary antibodies overnight at 4 °C. The cells were then stained with

fluorophore-conjugated secondary antibodies (Proteintech) in the dark,

and the nuclei were counterstained with DAPI (LEAGENE). Fluorescent

images were captured using a confocal microscope (Zeiss LSM 980).

Cycloheximide (CHX) assay

Cells were cultured in six-well plates and transfected with siRNA in

advance. After 48 h, the cells were treated with 25 μg/mL CHX (SigmaAldrich). Following treatment for 0, 6, 12, and 24 h, the cells were harvested

and lysed. Subsequently, the protein abundance was assessed by

performing western blot analysis using the indicated antibodies.

Protein-binding model

Molecular docking calculations were performed using MDockPP online

webserver (https://zougrouptoolkit.missouri.edu/MDockPP). The initial

protein structures of FASN and USP14 were predicted by AlphaFold2

and obtained from the Uniprot website (https://www.uniprot.org). The

protein identifier is AF-P54578-F1 and AF-P49327-F1 for FASN and USP14,

respectively. The FASN uiquitination sites were identified using hCKSAAP_UbSite for further analysis of docking conformation [33].

Xenograft tumour models

Male BALB/c nude mice (3 weeks old, n = 36) were obtained from

Guangdong Medical Laboratory Animal Center. The mice were randomly

assigned to different groups with no blinding. To establish murine xenograft

tumours, cells were subcutaneously injected into the right posterior flank of

the nude mice. Once the tumours reached a size of 150–200 mm3

, tumourbearing mice were either treated with IR (6 Gy) targeted at the tumour site

or administered sorafenib (60 mg/kg) intragastrically. The tumours development in the mice were monitored by measuring tumour volume,

calculated using the following formula: (width2 × length)/2. At the end of

the treatment, the mice were euthanised, and the tumour tissues were

collected for subsequent immunostaining analyses.

Clinical sample collection

Paraffin-embedded specimens from 36 newly diagnosed NPC patients and

20 recurrent NPC patients were generously provided by the Sun Yat‐sen

University Cancer Center, Guangzhou, China. The biopsy tissues were

pathologically confirmed as NPC. The collected samples were utilised for

staining experiments and analysis of clinical characteristics in correlation

with the available clinical data.

Immunohistochemical (IHC) staining

The tissue slides were deparaffinised in xylene and gradually rehydrated

using an alcohol gradient. Endogenous peroxidase activity was quenched

using 3% hydrogen peroxide, followed by antigen retrieval through

steaming with a 0.1 M sodium citrate solution (pH 6.0). Subsequently, the

sections were blocked with 5% BSA and incubated with the respective

primary antibodies overnight at 4 °C. Afterward, the slides were stained

with a secondary antibody and visualised using the GTVisionTM III

Detection System (GeneTech). Images were captured using the automated

microscope (Olympus BX63), and staining scores were evaluated based on

staining area and intensity.

Statistics

The data were analysed using GraphPad Prism (version 9.0) and are

presented as the mean ± SEM. Unless otherwise specified, each experiment

was independently conducted in triplicate, and the value of ‘n’ is indicated

in the figure legends. Student’s t test (two-tailed) was performed to

compare differences between two groups, while one-way analysis of

variance (ANOVA) was used for multiple group comparisons. Spearman’s

bivariate correlation analysis was utilised to calculate the correlations.

Kaplan–Meier analysis was employed to estimate overall survival and

progression-free survival.

RESULTS

IR induces lipid peroxidation and ferroptosis in NPC cells

Ferroptosis is directly induced by lipid peroxidation, primarily

originating from polyunsaturated fatty acid-containing phospholipids [10, 11]. We performed high-throughput targeted metabolomics of oxylipins to determine whether IR affects lipid

peroxidation in NPC cells. Compared with control treatment, IR

caused a significant increase in oxidised lipids in NPC cells

(Fig. 1a, b), particularly the oxidation products of arachidonic acid

and linoleic acid (Supplementary Fig. 1A). This observation

suggests a potential association between IR and the occurrence

of ferroptosis in NPC. Cell viability and clonogenic survival assays

on 5-8F, HONE1, and CNE2 cells indicated that IR-mediated cells

death could be partially restored by ferrostatin-1, an extensively

used ferroptosis antagonist (Fig. 1c, d). The multi-target single-hit

model showed that the cells treated with ferrostatin-1 exhibited

elevated values for D0, Dq, and SF2 compared to the control

group (Supplementary Table 4). This indicates that NPC cells

treated with ferrostatin-1 are less susceptible to IR and possess a

greater capacity for sublethal damage repair. The expression of

prostaglandin endoperoxidase synthase 2 (PTGS2), a marker gene

of ferroptosis [12], was upregulated upon IR, and this elevated

expression was reversed with ferrostatin-1 treatment (Fig. 1e).

C11-BODIPY 581/591 fluorescence staining revealed that lipid

peroxidation in NPC cells after IR significantly increased at 24 h

and stabilised when exposed to IR doses over 6 Gy (Supplementary Fig. 1B, C). The lipid peroxidation was significantly upregulated by 6 Gy IR exposure after 24 h in NPC cells (Fig. 1f and

Supplementary Fig. 1D). Using TEM, we observed that NPC cells

exposed to IR exhibited typical features of ferroptosis, including

mitochondrial shrinkage, increased membrane density, and

thickened cristae (Fig. 1g). Our results collectively indicate that

IR triggers ferroptosis in NPC cells.

GSTM3 facilitates ferroptosis upon IR exposure in vitro

We performed whole-transcriptome sequencing on 5-8 F and

CNE2 cells before and after IR and identified a series of

differentially expressed genes (Fig. 2a). Subsequently, we conducted Kyoto Encyclopaedia of Genes and Genomes (KEGG)

pathway analysis to determine the biological roles of these

Fig. 1 IR induces lipid peroxidation and ferroptosis in NPC cells. a Heatmap clustering of oxylipins in NPC cells from the control and IR

group. Columns: individual samples; rows: oxylipins; blue: low expression; red: high expression. b Volcano plot of the differential oxylipins

expression upon IR treatment. c, d Cell viability assays and clonogenic assays in 5-8 F, CNE2, and HONE1 cells pretreated with ferrostatin-1 or

DMEM for 24 h before exposure to 6 Gy IR. e NPC cell lines were pretreated with ferrostatin-1 or DMEM for 24 h, followed by exposure to 6 Gy

IR. The relative PTGS2 expression was assessed by qRT-PCR analysis. f, g The lipid peroxidation (f) and morphological changes of mitochondria

(g) in NPC cells with or without 6 Gy IR exposure. The lipid peroxidation levels were determined using C11-BODIPY 581/591 fluorescence

staining. The morphological changes of mitochondria were assessed via TEM. Scale bars: 500 nm. Data are presented as mean ± SEM, n = 3

independent repeats. p values were calculated using a two-tailed Student’s t test. *p < 0.05; **p < 0.01; ***p < 0.001.

Y. Chen et al.

4

British Journal of Cancer

differentially expressed genes. Our analysis revealed enrichment

of pathways related to lipid metabolism and amino acid

metabolism (Supplementary Fig. 2A). As the accumulation of

lipoperoxides is considered a hallmark of ferroptosis [10, 11], we

identified 107 differentially expressed genes shared by both 5-8F

and CNE2 cell lines, including five genes related to glutathione

metabolism and six genes related to lipid metabolism (Supplementary Fig. 2B). Among them, the expression of GSTM3 showed

5-8F CNE2

CON

1.2

1.0

0.8

0.6

0.4

0.2

0.0

1.2

1.0

0.8

0.6

0.4

0.2

0.0

0.1

1

0.01

IR

GSTM3

GSTM3

IR

GSTM3+IR

CON

IR+3-MA

CON

IR

GSTM3

Cell viability Cell viability

2.0

1.5

0.5

1.0

0.0

Relative PTGS2 expression Surviving fraction CNE2-R

Lipid oxidation level (% of control)

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Cell viability

Surviving fraction

– –

– –

+ +

+ +

– –

– –

+ +

+ +

– –

– –

+ +

+ +

IR

–2

–1

0

1

2

CON IR

–2

–1

0

1

2

–2

ANPEP CHAC1 GGT7 GSTM2 GSTM3 ACSBG1 ALOX15B APOE ELOVL3 PLA2G3 PLA2G4A

–1

0

1

2

3

4 bRelative mRNA expression

a

d

5-8F

ns

ns

ns

ns

HONE1 CNE2

5-8F HONE1 CNE2 5-8F HONE1 CNE2

5-8F

5-8F

HONE1 CNE2

5-8F

5-8F

HONE1

HONE1

CNE2

CNE2

e

25

55

– –

– –

+ +

+ +

– –

– –

+ +

+ +

– –

– –

+ +

+ + – –

– –

+ +

+ +

IR

GSTM3

GSTM3

-Tubulin

IR –+ –+ –+

25

55

25

55

IR

IR+Ferr-1

IR+Z-VAD-FMK

IR+Necro-1

GSTM3

CON

CNE2

CNE2-R

CNE2-R CNE2-R CNE2-R

IR+3-MA

CON

IR

IR+Ferr-1

IR+Z-VAD-FMK

IR+Necro-1

CON

GSTM3+IR

GSTM3

1.2

1.0

0.8

0.6

0.4

0.2

0.0

200

150

100

50

0

2.0

2.5

1.0

1.5

0.5

0.0

0.1

0.01

1

Cell viability

Lipid oxidation level Relative PTGS2 (% of control) expression

200

250

150

50

100

0

IR

GSTM3

– –

– –

+ +

+ +

– –

– –

+ +

+ +

– –

– –

+ +

+ +

CON+IR

CON

GSTM3+IR

GSTM3

CON+IR

GSTM3+IR

CON+IR

GSTM3+IR

CON+IR

CON

GSTM3

IR

IR

CON

CON

CON GSTM3

GSTM3 GSTM3+IR

500 nm

gh i

c

f

j kl

m

Y. Chen et al.

5

British Journal of Cancer

Fig. 2 IR-induced ferroptosis is regulated by GSTM3. a Heatmap of differentially expressed genes with or without 6 Gy IR exposure in 5-8F

and CNE2 cell lines. b qRT-PCR analysis of glutathione or lipid metabolism-associated genes after 6 Gy IR exposure. c Western blot analysis of

GSTM3 protein expression in NPC cell lines at 48 h after 6 Gy IR. d, e Cell viability assays and clonogenic assays in NPC cells that were

transiently transfected with GSTM3 or the empty vector plasmids followed by exposure to 6 Gy IR. f The cell viability assays detected the

effects of ferrostatin-1, 3-methyladenine, necrostatin-1, and Z-VAD-FMK in NPC cells with GSTM3 overexpression upon IR. g, h Relative PTGS2

mRNA expression (g) and lipid peroxidation levels (h) in empty vector- and GSTM3-overexpressing NPC cells following 6 Gy IR treatment.

i, j Cell viability assays and clonogenic assays in CNE2-R cells transiently transfected with GSTM3 or the empty vector plasmids followed by

exposure to 6 Gy IR. k, l PTGS2 mRNA expression (k) and lipid peroxidation (l) in GSTM3-overexpressing or empty vector-transfected CNE2-R

cells following 6 Gy IR. m TEM images of morphological changes in the mitochondria of CNE2-R cells. Scale bars: 500 nm. Data are presented as

mean ± SEM, n = 3 independent repeats. p values were calculated using the two-tailed Student’s t test. *p < 0.05; **p < 0.01; ***p < 0.001.

Implant

cells

Palpable

tumour

Day1

Day12

sacrifice

Day11

CON

GSTM3

CON + IR

GSTM3 + IR

Day10

IR 6 Gy

a

Day2 Day3

IR 6Gy every three days

Day4

10

2500

2000

1500

1000

500

2 4 6

Days

CON GSTM3 CON+IR

5-8F xenograft

GSTM3+IR

8 10 12

0.5

1.0

1.5

2.0

2.5

3.0

3.5

0.0 0

0

02468

Days

CON

GSTM3+

CON+IR

GSTM3+IR

10 12

ns

ns

ns

Mice weight(g)

Tumour weight (g) Tumour volume (mm3

)

15

20

25

30

CON

GSTM3+

CON+IR

GSTM3+IR

CON

GSTM3

CON+IR

GSTM3+IR

CON

GSTM3

CON+IR

GSTM3+IR

GSTM3+IR

CON

GSTM3

CON+IR

12

9

6

3 4-HNE staining GSTM3 staining 4-HNE staining score GSTM3 staining score 0

12

9

6

3

0

ns

b

cd e

f g

h

Fig. 3 GSTM3 potentiates the radiosensitivity of NPC cells in vivo. a Schematic diagram of treatment procedure in subcutaneous tumour

xenograft models. The 5-8F cells stably transfected with GSTM3 or empty vector lentivirus were used to construct subcutaneous tumour

xenograft models. b The weights of the mice from each group were measured every two days. c–e Macroscopic images (c), volume (d), and

weight (e) of the 5-8F xenograft tumours for each group (n = 4). f–h Representative IHC images (f) and staining scores of GSTM3 (g) and 4-HNE

(h) of subcutaneous tumours from each group. Scale bars: 50 μm/20 μm. Data are presented as mean ± SEM. p values were calculated using

two-tailed Student’s t test and two-way ANOVA. *p < 0.05; **p < 0.01; ***p < 0.001.

Y. Chen et al.

6

British Journal of Cancer

the most significant upregulation following IR exposure, as

determined by qRT-PCR assay (Fig. 2b). The results revealed that

IR substantially increased both mRNA and protein expression of

GSTM3 in NPC cell lines (Fig. 2c and Supplementary Fig. 2C, D).

GSTM3 expression gradually stabilised following increased IR

doses (Supplementary Fig. 2E). These results indicate that GSTM3

may serve as a potential modulator of radiosensitivity in NPC

patients.

To further explore the potential role of GSTM3, we constructed

NPC cells with overexpression GSTM3 (Supplementary Fig. 2F). Cell

Vector

GSTM3

Vector

Vector

USP14

Vector

USP14

GSTM3

GSTM3

Flag-GSTM3

plasmid Transfection

a b

d

5-8F cells

IP (Anti-Flag)

LC-MS analysis

USP14 IP : Flag Input CNE2 5-8F

Eluant

USP14

5-8F CNE2

5-8F

5-8F

HONE1 CNE2

CNE2

5-8F CNE2

5-8F

5-8F 5-8F

CNE2

siNC

0 h 6 h 12 h 24 h 0 h 6 h 12 h 24 h

0 h 6 h 12 h 24 h 0 h 6 h 12 h 24 h

250

70

55

250

70

55

250

70

55

250

70

55

CHX

siUSP14

MG-132

siUSP14

MG-132

HA-Ub

FASN

siUSP14

CNE2 CNE2

–

– –

+ +

+

–

– –

+ +

+

siNC siUSP14

siNC siGSTM3-1 siGSTM3-2 siNC siGSTM3-1 siGSTM3-2

5-8F HONE1

55

ns ns ns Vector

USP14

70

25

55

250

70

0.5

0.0

1.0

1.5

55

70

25

55

70

25

-Tubulin

GSTM3

USP14

-Tubulin

USP14

FASN

USP14 FASN DAPI Merge

10 M

-Tubulin

FASN

USP14

FASN 250

70

250

70

55

Flag Flag-USP14 Flag Flag-USP14

USP14

-Tubulin

USP14

-Tubulin

CHX

FASN

USP14

-Tubulin

FASN

USP14

-Tubulin

siUSP14

MG-132

FASN

USP14

-Tubulin

FASN

FASN

IB

USP14

-Tubulin

Relative mRNA of FASN

–

+

+

+

+

+

–

+

+

+

+

+

IP : HA Input

5-8F 5-8F

1.2

1.0

0.8

0.6

0.4

0.2

0.0

3.0

2.5

2.0

1.5

1.0

0.5

0.0

2.5 250

200

150

100

50

0

200

150

100

50 Lipid oxidation level (% of control) Lipid oxidation level (% of control)

0

2.0

1.5

1.0

0.5

0.0

siFASN-1

siFASN-2

siFC 5-8F

siFASN-1

siFASN-2

siFC

5-8F

siFASN-1

siFASN-2

siFC –

– –

+ +

+

CNE2 CNE2 CNE2

CON IR

CON IR CON IR CON IR

CON IR

CNE2

CON IR

–

– –

+ +

+

siUSP14

GSTM3

GSTM3

FASN

USP14

-Tubulin

siUSP14

GSTM3

GSTM3

FASN

USP14

-Tubulin

250

70

25

55

250

70

25

55

Cell viability

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Cell viability

Relative PTGS2 expression Relative PTGS2 expression

f g

e

c

h ij

k lmn

siFASN-1

siFASN-2

siFC

siFASN-1

siFASN-2

siFC

siFASN-1

siFASN-2

siFC

250

250

70

55

Y. Chen et al.

7

British Journal of Cancer

viability and clonogenic survival assays showed that GSTM3

overexpression significantly promoted IR-induced cell death

(Fig. 2d, e). Remarkably, the enhancement of radiosensitivity by

GSTM3 was nearly nullified upon treatment with the ferroptosis

inhibitor ferrostatin-1, while the utilisation of the autophagy

inhibitor 3-methyladenine, the necroptosis inhibitor necrostatin-1,

or the apoptosis inhibitor Z-VAD-FMK did not produce a

comparable effect (Fig. 2f). Furthermore, GSTM3 increased the

IR-induced PTGS2 expression and lipid peroxidation (Fig. 2g, h and

Supplementary Fig. 2G). Conversely, GSTM3 silencing in 5-8F and

HONE1 cell lines moderately attenuated IR-induced cell death

(Supplementary Fig. 3A, B). IR-induced PTGS2 expression and lipid

peroxidation were severely reduced by the silencing of GSTM3

(Supplementary Fig. 3C, D). Collectively, these findings elucidate

that GSTM3 is critical for modulating NPC radiosensitivity

predominantly by promoting IR-induced ferroptosis.

Subsequently, we generated a radiation-resistant NPC cell line

termed CNE2-R and validated its increased resistance to IR in

comparison to the parental CNE2 cells (Supplementary Fig. 4A–C).

Gamma-H2AX (γ-H2AX), a hallmark of DNA double-stranded

breaks, is a critical marker for effectively detecting DNA damage

and repair response [34]. Western blot analysis showed a timedependent increase in γ-H2AX expression following IR in CNE2

cells, whereas CNE2-R cells exhibited a comparatively lower level

of γ-H2AX expression (Supplementary Fig. 4D), indicating that

CNE2-R cells were less responsive to IR-induced DNA damage and

subsequent repair response. Additionally, treatment of CNE2-R

cells with ferrostatin-1 did not cause any restoration of cell death

induced by IR (Supplementary Fig. 4E, F), and IR did not increase

the PTGS2 expression and lipid peroxidation in CNE2-R cells

(Supplementary Fig. 4G, H). Furthermore, we observed that the

expression of GSTM3 in CNE2-R cells remained unaltered despite

exposure to different doses of IR (Supplementary Fig. 4I).

To determine whether GSTM3 mediates IR-induced ferroptosis

to alleviate radioresistance in CNE2-R cells, we overexpressed

GSTM3 in CNE2-R cells (Supplementary Fig. 4J). GSTM3 overexpression caused decreased cell viability and clonogenic survival

upon exposure to IR (Fig. 2i, j). Additionally, upon IR treatment,

PTGS2 expression and lipid peroxidation were significantly

enhanced in CNE2-R cells overexpressing GSTM3 (Fig. 2k, l and

Supplementary Fig. 4K). TEM revealed that IR induced subtle

morphological changes in the mitochondria of CNE2-R cells.

However, upon GSTM3 overexpression, these mitochondrial

changes became more pronounced after IR exposure, characterised by a shrunken shape, increased membrane density, and

thickened cristae (Fig. 2m). Overall, our data strongly suggest that

GSTM3 effectively alleviates radioresistance by potentiating IRinduced ferroptosis.

GSTM3 promotes IR-mediated ferroptosis and NPC

radiosensitivity in vivo

To investigate the impact of GSTM3 on IR-induced ferroptosis

in vivo, we subcutaneously injected 5-8F cells with stable GSTM3

overexpression into nude mice, leading to the formation of palpable

tumours. Subsequently, the mice bearing xenograft tumours were

subjected to regular IR treatments (Fig. 3a). On day 12, the mice

were euthanised, and the tumours were dissected for volume

measurement and IHC staining. The weights of the mice remained

stable throughout the entire treatment period (Fig. 3b). Compared

with the control groups, GSTM3 overexpression alone did not

exhibit any impact on tumour growth in xenograft models, whereas

IR effectively suppressed tumour growth (Fig. 3c–e). Notably,

compared to the group treated with IR alone, GSTM3 overexpression

combined with IR treatment resulted in a substantial reduction in

tumour size and weight (Fig. 3c–e). 4-hydroxy-2-noneal (4-HNE) acts

as a ferroptosis marker reflecting the level of lipid peroxidation [35].

IHC staining revealed that IR moderately increased the abundance of

GSTM3 and 4-HNE (Fig. 3f–h). Moreover, GSTM3 overexpression

combined with IR treatment led to a significantly higher level of the

4-HNE signal (Fig. 3f–h). Collectively, these results suggest that

GSTM3 enhances IR-mediated ferroptosis and improves radiosensitivity in NPC.

GSTM3 stabilises USP14/FASN axis to promote IR-induced

ferroptosis

To further explore the mechanism underlying how GSTM3

facilitates IR-induced ferroptosis, we conducted LC-MS analysis

and identified the deubiquitinase USP14 as a potential target of

GSTM3 (Fig. 4a and Supplementary Fig. 5A). Western blot analysis

indicated that GSTM3 overexpression effectively increased the

protein expression of USP14 (Fig. 4b and Supplementary Fig. 5B),

whereas GSTM3 silencing decreased the expression of USP14

(Fig. 4c and Supplementary Fig. 5C). It was proposed that USP14

regulates lipid and carbohydrate metabolism in hepatosteatosis

by stabilising FASN, a key lipogenic enzyme [36]. We observed

that the protein expression of FASN was correspondingly

increased in NPC cells with overexpressing USP14 (Fig. 4d and

Supplementary Fig. 5D), whereas silencing USP14 reduced the

protein expression of FASN (Supplementary Fig. 5E). However, the

mRNA expression of FASN was independent of USP14 overexpression (Fig. 4e), suggesting that USP14 modulates FASN at the

posttranscriptional level. Co-IP assays revealed an interaction

between USP14 and FASN (Fig. 4f). IF analysis confirmed the colocalisation of USP14 and FASN in the cytoplasm (Fig. 4g). USP14

consists of an N-terminal ubiquitin-like domain (UBL) and a

C-terminal catalytic domain (CAT) [37]. FASN is a homodimer

comprising seven functional domains [38]. Employing the proteinbinding model [33], we observed that USP14CAT was predicted to

interact with FASNTE and FASNMAT (Supplementary Fig. 5F).

Furthermore, CHX analysis confirmed that USP14 silencing

significantly promoted the degradation of endogenous FASN

(Fig. 4h and Supplementary Fig. 5G), suggesting that USP14 can

lengthen the half-life of FASN protein. As USP14 is a major

regulator of the proteasome and possesses proteasomeassociated deubiquitination activity [39], we treated NPC cells

with MG132, a proteasome suppressor that impedes ubiquitin

degradation. The results showed that silencing USP14-mediated

destabilisation of FASN was partly reversed by MG132 (Fig. 4i and

Fig. 4 GSTM3 promotes IR-induced ferroptosis partly via the USP14/FASN axis. a LC-MS analysis identified USP14 as a potential target of

GSTM3. b, c Western blot analysis of USP14 in NPC cell lines with overexpressed GSTM3 or silencing GSTM3. d, e The protein expression or

mRNA level of FASN in NPC cells transiently transfected with USP14 or the empty vector plasmids. f Co-immunoprecipitation assays with antiFlag antibodies in 5-8F and CNE2 cells revealed that USP14 interacts with FASN in vitro. g Immunofluorescence staining revealed the colocalisation of endogenous USP14 (green) and FASN (red) in the cytoplasm. Scale bars: 10 μm. h Protein stability of FASN was determined by

CHX treatment analysis in NPC cells transfected with siUSP14 or control siRNA. i The effect of MG132 in siUSP14 or siNC NPC cell lines. The

protein expression of FASN was measured via western blot analysis. j NPC cells were transfected with siUSP14 or siNC, as well as HA-Ub

plasmid for 48 h. Lysates from cells were immunoprecipitated with anti-HA. Western blot analysis with the indicated antibody was conducted

to analyse ubiquitination levels of FASN. k Western blot analysis in NPC cells co-transfected with either GSTM3 or the empty vector plasmids

alongside siUSP14 or siNC. l–n 5-8F and CNE2 cells were transfected with either siFASN or control siRNA. The cell viability (l), the relative PTGS2

mRNA level (m), and lipid peroxidation level (n) were assessed in respective cells after exposure to 6 Gy IR. Data are presented as mean ± SEM,

n = 3 independent repeats. p values were calculated using the two-tailed Student’s t test. ***p < 0.001.

Y. Chen et al.

8

British Journal of Cancer

Supplementary Fig. 5H). Ubiquitination assays indicated that the

polyubiquitination of FASN was increased by USP14 silencing in

NPC cells (Fig. 4j), demonstrating that USP14 stabilises the FASN

protein by inhibiting its ubiquitin-proteasome degradation in NPC

cells. Importantly, western blot analysis showed that FASN and

USP14 were upregulated by GSTM3 overexpression, whereas FASN

expression was almost completely restored by USP14 silencing

(Fig. 4k and Supplementary Fig. 5I). These results indicate that

GSTM3 regulates USP14 expression to inhibit the polyubiquitination and subsequent degradation of FASN.

Moreover, cell viability assay showed that FASN silencing resulted

in significant restoration of cell viability under IR exposure (Fig. 4l and

Supplementary Fig. 6A, B). Furthermore, FASN silencing mitigated IRinduced PTGS2 expression and lipid peroxidation (Fig. 4m, n and

Supplementary Fig. 6C), suggesting that FASN mediates IR-induced

ferroptosis to improve radiosensitivity in NPC. Collectively, our

findings indicate that GSTM3 promotes IR-induced ferroptosis and

enhances radiosensitivity through USP14/FASN axis in NPC.

GSTM3 facilitates IR-induced ferroptosis by suppressing GPX4

To investigate the possibility of additional mechanisms underlying

IR-induced ferroptosis promoted by GSTM3, we examined the

expressions of several key genes (GPX4, ACSL3, SLC7A11, TF, FTL)

involved in the ferroptosis pathways [11, 12] in NPC cells with

GSTM3 overexpression. Western blot analysis revealed that the

overexpression of GSTM3 reduced the expression of GPX4

(Supplementary Fig. 7A), a glutathione peroxidase possessing

potent antioxidant activity [40]. Additionally, Co-IP assays revealed

an interaction between GSTM3 and GPX4 (Supplementary Fig. 7B).

IF staining demonstrated the co-localisation of GSTM3 and GPX4

in the cytoplasm (Supplementary Fig. 7C), indicating that GSTM3

interacts with GPX4 and suppresses its expression. To investigate

the potential involvement of GPX4 in IR-induced ferroptosis, we

conducted functional experiments in NPC cells overexpressing

GPX4 (Supplementary Fig. 7D). Remarkably, the overexpression of

GPX4 significantly reversed IR-induced cell death and notably

reduced the levels of PTGS2 and lipid peroxidation following IR

treatment (Supplementary Fig. 7E–G). These results suggest that

GPX4 acts as a downstream target of GSTM3, mediating IRinduced ferroptosis in NPC. Furthermore, in the subcutaneous

tumour xenograft model, the expression of USP14 and FASN were

increased, while the protein levels of GPX4 were decreased in the

group with GSTM3 overexpression (Supplementary Fig. 8A). The

schematic diagram illustrating the mechanisms by which GSTM3

enhances IR-induced ferroptosis in NPC was presented in

Supplementary Fig. 8B.

FINs and IR synergistically trigger ferroptosis and sensitise

cancer cells to radiotherapy

Ferroptosis can be induced by different types of FINs [16, 17],

prompting us to investigate whether the combined treatment of

FINs and IR could synergistically potentiate ferroptosis and

enhance NPC radiosensitivity. In this study, the FINs sorafenib

and FIN56, which inhibit SLC7A11 and GPX4 respectively, were

used to explore the potential role in NPC. Individual treatment

with FINs or IR suppressed the NPC cells viability and promoted

PTGS2 expression and lipid peroxidation level (Fig. 5a–c). Notably,

the combined treatment with FINs and IR resulted in a dramatic

increase in NPC cells death, the PTGS2 expression, and lipid

peroxidation (Fig. 5a–c). Considering the selectivity of sorafenib as

a ferroptosis inducer for certain tumour cell lines [41], we

conducted further investigations to assess its cytotoxic effects in

NPC cell lines. The cell viability assays revealed that sorafenib

induced cell death in 5-8F and CNE2, while ferrostatin-1 rescued

the lethal effects upon sorafenib treatment (Supplementary

Fig. 9A). C11-BODIPY fluorescence staining indicated an elevation

in lipid peroxidation levels induced by sorafenib, which could be

counteracted by ferrostatin-1 (Supplementary Fig. 9B). Moreover,

the synergistic effect of sorafenib and IR was mitigated in the

presence of ferrostatin-1 treatment (Supplementary Fig. 9A, B).

Given that sorafenib is a multiple-target tyrosine kinase inhibitor,

we further investigated the synergy between IR and downregulation of SLC7A11. The results showed that the downregulation of SLC7A11 significantly promoted IR-induced cell

death and increased lipid peroxidation following IR treatment

(Supplementary Fig. 9C–E).

Subsequently, we generated subcutaneous tumour xenograft

models to assess the combined effect of sorafenib and IR on NPC

radiosensitivity in vivo (Fig. 5d). The results indicated that the

individual treatment with sorafenib or IR reduced the volume and

weight of tumours (Fig. 5e–g). Significantly, the combined

treatment of sorafenib and IR exhibited substantial suppression

of xenograft growth (Fig. 5e–g). Sorafenib and IR increased the

expression of 4-HNE, a hallmark of ferroptosis, and decreased the

expression of proliferating cell nuclear antigen (PCNA), a

biomarker of proliferation (Fig. 5h). Notably, the combination of

sorafenib with IR treatment further enhanced 4-HNE expression

and suppressed PCNA expression (Fig. 5h). These findings reveal

that the combined treatment of FINs and IR synergistically

potentiates ferroptosis in NPC cells, resulting in a significant

sensitisation of cancer cells to radiotherapy.

Low expression of GSTM3 is correlated with tumour relapse

and poor prognosis in NPC

To evaluate the clinical significance of GSTM3 in radiotherapy

response and prognosis, we collected tumour tissues from 36

newly diagnosed NPC patients and 20 recurrent NPC patients.

Remarkably, the newly diagnosed NPC samples exhibited strongly

positive staining for GSTM3 and 4-HNE (Fig. 6a), while the

recurrent tumour tissues showed relatively negative expression of

GSTM3 and 4-HNE (Fig. 6b). The association between the clinical

characteristics and the expression of GSTM3 and 4-HNE in NPC

patients is presented in Supplementary Tables 5 and 6,

respectively. These results indicated that tumour relapse was

associated with weak expression of GSTM3 and 4-HNE in NPC

patients (Fig. 6c and Supplementary Fig. 10A, B). Moreover, there

is a positive association between GSTM3 and 4-HNE staining

(Fig. 6d). A moderate or strong positive dual staining signal of

GSTM3 and 4-HNE was correlated with low incidence rates of NPC

recurrence (Fig. 6e). Kaplan–Meier analysis revealed that a low

expression of either GSTM3 or 4-HNE was associated with poor

overall survival and progression-free survival in NPC patients

(Fig. 6f, g). Furthermore, the simultaneous low expressions of both

GSTM3 and 4-HNE conferred a higher risk of disease progression

and mortality (Fig. 6f, g). Based on The Cancer Genome Atlas, we

found that high GSTM3 expression was correlated with longer

disease-free survival (Supplementary Fig. 10C). These findings

suggest that GSTM3 contributes to the radiotherapy response

associated with ferroptosis, predicting a favourable prognosis in

NPC patients.

DISCUSSION

Radiotherapy is a standard treatment modality for patients

diagnosed with NPC [2]. However, some patients suffer from

residual neoplasms and local recurrence due to radioresistance [5].

NPC radioresistance is attributed to various regulatory mechanisms, including DNA damage response signalling, cancer stem cell

phenotype, abnormal cell-cycle progression, hypoxic properties,

and the tumour immune microenvironment [42–45]. However, our

current understanding of these mechanisms has not fully

addressed the challenges of radioresistance and treatment failure.

Consequently, it is crucial to explore the underlying mechanisms

of radioresistance to identify potential strategies for improving

therapeutic response.

Y. Chen et al.

9

British Journal of Cancer

Ferroptosis plays a critical role in mediating radiation response in

hepatocellular carcinoma, lung cancer, and melanoma [12–14].

Studies have reported that Epstein–Barr virus (EBV) infectioninduced GPX4 can enhance chemoresistance and promote tumour

progression by inhibiting ferroptosis [46]. Cephalosporin,

itraconazole, and cucurbitacin B have been shown to trigger

ferroptosis to exert potential antitumour effects in NPC [20–23].

However, the impact of ferroptosis on NPC radiosensitivity remains

unclear. Ferroptosis is characterised by the excessive peroxidation

of polyunsaturated fatty acid-containing phospholipids [11]. In this

1.2

a

1.0

0.8

0.6

0.4

0.2

0.0

Cell viability Tumor volume (mm3

)

4-HNE staining PCNA staining Tumor weight (g) 1.2

d

1.0

0.8

0.6

0.4

0.2

0.0

Cell viability

Relative PTGS2 expression Lipid oxidation level (% of control) Lipid oxidation level (% of control) Relative PTGS2 expression

CON

IR

CON

IR CON

IR

5-8F

CON

IR

5-8F

CON

IR

5-8F

CON

IR

5-8F

FIN56 Sora

0

1

2

3

4

5

0

1

2

3

4

6

5

100

200

300

0

100

200

300

0

IR

IR

Sora + IR

IR

Sora + IR

Sora + IR

5-8F Xenograft

Sora

Sora

Sora

CON

Tumour

bearing

CON

IR

Sora + IR

Sora

3000 CON

2000

1000

0

0 2 4 6 8 10

Days

3.0

2.5

2.0

1.5

1.0

0.5

0.0

Sacrifice

CON

0 123 4 5 6 7 8 9 10

CNE2 CNE2

DMSO DMSO FIN56 Sora DMSO FIN56 Sora

DMSO FIN56 Sora DMSO FIN56 Sora DMSO FIN56 Sora

Sora 25 mg/kg

Xenograft

model

IR 6 Gy

IR

Sora + IR

CON

Sora

f h

e

g

b c

Fig. 5 FINs sensitise cancer cells to radiotherapy in vitro and in vivo. a–c 5-8F and CNE2 cell lines were pretreated with FIN56, sorafenib, or

DMEM for 24 h before 6 Gy IR exposure. The cell viability (a), relative PTGS2 expression (b), and lipid peroxidation levels (c) were measured in

the respective cell lines. d The therapeutic modality in the mice with 5-8F subcutaneously xenografted tumours. The mice with xenografts

were treated with 6 Gy IR or intragastric administration of sorafenib. e–g The size (e), volume (f), and weight (g) of the xenograft tumours in

control, sorafenib, IR, or combination therapy groups (n = 5). h IHC staining images of 4-HNE and PCNA of xenograft tumours from each

group. Scale bars: 50 μm/20 μm. Data are presented as mean ± SEM. p values were calculated using two-tailed Student’s t test and two-way

ANOVA. *p < 0.05; **p < 0.01; ***p < 0.001.

Y. Chen et al.

10

British Journal of Cancer

study, we conducted a targeted metabolomic analysis of oxylipins

in NPC cells and found that IR resulted in an increase in lipid

peroxidation, specifically the oxidation products of arachidonic acid

and linoleic acid. Importantly, we demonstrated that IR-induced cell

death could be rescued by the ferroptosis inhibitor ferrostatin-1.

Furthermore, IR exposure resulted in an increase of PTGS2

expression, accumulation of lipid peroxidation, and characteristic

morphological changes in NPC cells, collectively suggesting that

ferroptosis potentially plays a crucial role in regulating the

sensitivity of NPC to radiotherapy.

GSTM3 participates in the intricate processes of tumourigenesis

and the progression in various malignancies. GSTM3 reportedly

Patient N1

GSTM3 staining

Patient N2

Newly diagnosed NPC patient (n = 36)

Patient N3 Patient N4 Patient R1 Patient R2 Patient R3 Patient R4

Recurrent NPC patient (n = 20)

4-HNE staining

4-HNE staining score 4HNE staining score Frequency (%) GSTM3 staining score GSTM3 staining 4-HNE staining

12

9

6

3

0

100

Overall survival (%) Overall survival (%) Progression-free survival (%)

Overall survival (%)

50

0

0 20 40

GSTM3low

GSTM3high

60

Survival time (months)

Log-rank p = 0.0016

80 100

Newly dignosed NPC

Recurrent NPC

Newly dignosed NPC

Recurrent NPC

12

9

6

3

0

12

12

15

15

r = 0.6141

p < 0.0001

GSTM3 Staining score

9

9

6

6

3

3

0

0 0

GSTM3

–/

–/+ –/+

++/+++

++/+++

++/+++

++/+++

4-HNE –/+

5

15

20

25

30

35

10

Newly diagnosed NPC

Recurrent NPC

0 20 40

4-HNElow GSTM3low/4-HNElow

GSTM3low/4-HNEhigh & GSTM3high/4-HNElow

GSTM3high/4-HNE 4-HNE high high

60

Survival time (months)

80 100

100

50

0

100

50

0

0 20 40 60

Survival time (months)

80 100

100

50

0

0 20 40 60

Survival time (months)

80 100

100

50

0

0 20 40 60

Survival time (months)

80 100

Progression-free survival (%) Progression-free survival (%)

Log-rank p = 0.0004 Log-rank p < 0.0001

100

50

0

0 20 40 60

Survival time (months)

80 100

Log-rank p = 0.0399 Log-rank p = 0.0309 Log-rank p = 0.0459

GSTM3low

GSTM3high

4-HNElow

4-HNEhigh GSTM3low/4-HNElow

GSTM3low/4-HNEhigh &

GSTM3high/4-HNElow

GSTM3high/4-HNEhigh

a b

c de

f

g

Fig. 6 Low expression of GSTM3 is correlated with tumour relapse and poor prognosis in NPC. a, b Representative IHC images of GSTM3

and 4-HNE staining in patients with newly-diagnosed NPC (a) or recurrent NPC (b). Scale bars: 50 μm/20 μm (inset). c IHC scoring of GSTM3 and

4-HNE staining in patients with newly-diagnosed NPC or recurrent NPC. d Correlations between the staining scores of GSTM3 and 4-HNE.

e The proportion of recurrence status in patients with GSTM3 and 4-HNE expression as detected by IHC. f, g Kaplan–Meier analysis of overall

survival and progression-free survival grouping by GSTM3 and 4-HNE expression levels. Data are presented as mean ± SEM. p values were

calculated using two-tailed Student’s t test. ****p < 0.0001.

Y. Chen et al.

11

British Journal of Cancer

predicts high susceptibility, tumour malignant behaviours, and

poor prognosis in some cancer types [28–32]. However, GSTM3

acts as a tumour suppressor to inhibit tumourigenesis in gastric

cancer, alleviate the aggressiveness in renal cell carcinoma, and

alter the malignant metabolic pattern in pancreatic cancer [25,

26, 47]. Moreover, GSTM3 reverses the radioresistance through

cell cycle arrest and apoptosis facilitation in hepatocellular

carcinoma [27]. However, the functions and mechanisms of

GSTM3 in NPC have not been investigated. Our study revealed

that GSTM3 was upregulated by IR and sensitised NPC cells to

radiotherapy by potentiating IR-mediated ferroptosis. Additionally,

GSTM3 alleviated the radioresistance in radiation-resistant NPC

cells. These findings suggest that GSTM3 holds promise as a

potential biomarker for promoting radiotherapy sensitivity in NPC.

While the prognostic value of GSTM3 has been reported in various

tumours, its potential as a prognostic indicator in NPC remains

uncertain. In our study, we observed that a low GSTM3 expression

conferred a higher risk of locoregional recurrence and predicted

poor overall survival and progression-free survival in NPC patients.

Our findings indicate that GSTM3 serves as a prognostic indicator

for NPC, laying the foundation for exploring potential treatment

strategies for patients with radiation-resistant or recurrent NPC.

USP14, a major deubiquitinase reversibly associated with the

proteasome, participates in IR-induced DNA double-strand breaks

repair via non-homologous end joining [48]. Moreover, USP14 is

involved in regulating ferroptosis, autophagy, amino acid metabolism, and immune suppression to mediate tumour progression

and treatment response [49, 50]. In this study, we found that

GSTM3 stabilised the expression of deubiquitinase USP14, thereby

inhibiting the ubiquitination and subsequent degradation of

FASN. As a pivotal enzyme involved in the lipid biosynthesis

pathway, FASN may supply polyunsaturated fatty-acid for the

production of lipid peroxidase. O-GlcNAcylation enhances the

transcriptional activity of FASN to facilitate ferroptosis in

mesenchymal pancreatic cancer [51]. FASN is related to Alzheimer’s disease-related toxicity that modulates lipid peroxidation

and induces ferroptosis [52]. However, FASN remodels oxidised

phospholipids to escape ferroptosis in KRAS-mutant lung cancer

[53]. In our study, we demonstrated that GSTM3 regulates USP14/

FASN axis to potentiate IR-induced ferroptosis in NPC, presumably

by enhancing the synthesis of lipid peroxidation.

As a glutathione peroxidase, GPX4 converts lipid hydroperoxides into lipid alcohols, ultimately eliminating lipid peroxidation

[40]. GPX4 inhibition-mediated ferroptosis is essential for the

radiosensitivity of breast cancer and hepatocellular carcinoma

[54, 55]. Additionally, ferroptosis resulting from GPX4 deficiency is

associated with antitumour immunity, malignant biological

properties, and platinum drugs resistance [56, 57]. EBV infectioninduced GPX4 reduces the sensitivity of cells to ferroptosis via

p62-Keap1-NRF2 signalling pathway, leading to chemoresistance

and tumour progression [46]. Cucurbitacin B and lupeol initiate

the mechanism of ferroptosis in NPC by downregulating the

expression of GPX4 [23, 58]. Our findings reveal that GPX4 serves

as a downstream effector of GSTM3 to regulate IR-induced

ferroptosis and NPC radiosensitivity.

FINs exert significant effects on improving radiotherapy

sensitivity and enhancing the immunotherapy response [12, 17].

Numerous studies have demonstrated that sorafenib showed

potent induction of ferroptosis in various cancers, unravelling

novel mechanisms through which sorafenib induced this process

[35, 59, 60]. However, the status of sorafenib as one of the FIN has

recently been challenged [41]. The selectivity of sorafenib as a

ferroptosis inducer in specific cell lines and the intricate molecular

mechanisms involved remain largely unknown. Intriguingly, we

found that sorafenib combined with IR synergistically triggered

ferroptosis and exhibited significant radiosensitising effects in

NPC. These findings hold promise in overcoming the challenge of

radiotherapy resistance in NPC and may provide clinicians with

new strategies to improve treatment outcomes. Although

SLC7A11 was not regulated by GSTM3, the targeting of SLC7A11

by sorafenib substantially promoted radiotherapy sensitisation,

implying that other critical pathways parallel to GSTM3 might

mediate IR-induced ferroptosis.

Overall, this study revealed the significant role of GSTM3 in

IR-induced ferroptosis and radiotherapy sensitivity in NPC. It

highlighted the involvement of GSTM3 in stabilising the USP14/

FASN axis and targeting GPX4 as key mechanisms underlying

these processes. Combining IR treatment with ferroptosis inducers

synergistically improved NPC radiosensitivity and suppressed

tumour growth. Notably, a decrease in GSTM3 abundance

predicted tumour relapse and poor prognosis. These findings

provide valuable insights for the development of promising

treatment strategies targeting radiation-resistant or recurrent NPC.

DATA AVAILABILITY

The datasets used and/or analysed in the current study are available from the

corresponding author on reasonable request.

REFERENCES

1. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global

cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209–49.

2. Chen YP, Chan ATC, Le QT, Blanchard P, Sun Y, Ma J. Nasopharyngeal carcinoma.

Lancet. 2019;394:64–80.

3. Chen YP, Ismaila N, Chua M, Colevas AD, Haddad R, Huang SH, et al. Chemotherapy in combination with radiotherapy for definitive-intent treatment of

stage II-IVA nasopharyngeal carcinoma: CSCO and ASCO guideline. J Clin Oncol.

2021;39:840–59.

4. Lee AW, Ma BB, Ng WT, Chan AT. Management of nasopharyngeal carcinoma:

current practice and future perspective. J Clin Oncol. 2015;33:3356–64.

5. Luftig M. Heavy LIFting: tumor promotion and radioresistance in NPC. J Clin

Invest. 2013;123:4999–5001.

6. Zhou X, Lin Y, Chen Y, Wang L, Peng X, Liao J, et al. Epstein-Barr virus (EBV)

encoded microRNA BART8-3p drives radioresistance-associated metastasis in

nasopharyngeal carcinoma. J Cell Physiol. 2021;236:6457–71.

7. Krause M, Dubrovska A, Linge A, Baumann M. Cancer stem cells: radioresistance,

prediction of radiotherapy outcome and specific targets for combined treatments. Adv Drug Deliv Rev. 2017;109:63–73.

8. Lin Y, Zhou X, Yang K, Chen Y, Wang L, Luo W, et al. Protein tyrosine phosphatase

receptor type D gene promotes radiosensitivity via STAT3 dephosphorylation in

nasopharyngeal carcinoma. Oncogene. 2021;40:3101–17.

9. Lee SY, Jeong EK, Ju MK, Jeon HM, Kim MY, Kim CH, et al. Induction of metastasis,

cancer stem cell phenotype, and oncogenic metabolism in cancer cells by

ionizing radiation. Mol Cancer. 2017;16:10.

10. Jiang X, Stockwell BR, Conrad M. Ferroptosis: mechanisms, biology and role in

disease. Nat Rev Mol Cell Biol. 2021;22:266–82.

11. Stockwell BR, Friedmann AJ, Bayir H, Bush AI, Conrad M, Dixon SJ, et al. Ferroptosis: a regulated cell death nexus linking metabolism, redox biology, and disease. Cell. 2017;171:273–85.

12. Lei G, Zhang Y, Koppula P, Liu X, Zhang J, Lin SH, et al. The role of ferroptosis in

ionizing radiation-induced cell death and tumor suppression. Cell Res.

2020;30:146–62.

13. Lang X, Green MD, Wang W, Yu J, Choi JE, Jiang L, et al. Radiotherapy and

immunotherapy promote tumoral lipid oxidation and ferroptosis via synergistic

repression of SLC7A11. Cancer Discov. 2019;9:1673–85.

14. Yang M, Wu X, Hu J, Wang Y, Wang Y, Zhang L, et al. COMMD10 inhibits

HIF1alpha/CP loop to enhance ferroptosis and radiosensitivity by disrupting CuFe balance in hepatocellular carcinoma. J Hepatol. 2022;76:1138–50.

15. Chen X, Kang R, Kroemer G, Tang D. Broadening horizons: the role of ferroptosis

in cancer. Nat Rev Clin Oncol. 2021;18:280–96.

16. Wang W, Green M, Choi JE, Gijon M, Kennedy PD, Johnson JK, et al. CD8(+) T cells

regulate tumour ferroptosis during cancer immunotherapy. Nature.

2019;569:270–4.

17. Ghoochani A, Hsu EC, Aslan M, Rice MA, Nguyen HM, Brooks JD, et al. Ferroptosis

inducers are a novel therapeutic approach for advanced prostate cancer. Cancer

Res. 2021;81:1583–94.

18. Koppula P, Zhuang L, Gan B. Cystine transporter SLC7A11/xCT in cancer: ferroptosis, nutrient dependency, and cancer therapy. Protein Cell. 2021;12:

599–620.

Y. Chen et al.

12

British Journal of Cancer

19. Liu Y, Wan Y, Jiang Y, Zhang L, Cheng W. GPX4: the hub of lipid oxidation,

ferroptosis, disease and treatment. Biochim Biophys Acta Rev Cancer.

2023;1878:188890.

20. Xu Y, Wang Q, Li X, Chen Y, Xu G. Itraconazole attenuates the stemness of

nasopharyngeal carcinoma cells via triggering ferroptosis. Environ Toxicol.

2021;36:257–66.

21. He X, Yao Q, Fan D, Duan L, You Y, Liang W, et al. Cephalosporin antibiotics

specifically and selectively target nasopharyngeal carcinoma through HMOX1-

induced ferroptosis. Life Sci. 2021;277:119457.

22. Li Y, Chen F, Chen J, Chan S, He Y, Liu W, et al. Disulfiram/Copper induces

antitumor activity against both nasopharyngeal cancer cells and cancerassociated fibroblasts through ROS/MAPK and ferroptosis pathways. Cancers.

2020;12:138.

23. Huang S, Cao B, Zhang J, Feng Y, Wang L, Chen X, et al. Induction of ferroptosis in

human nasopharyngeal cancer cells by cucurbitacin B: molecular mechanism and

therapeutic potential. Cell Death Dis. 2021;12:237.

24. Liu X, Campbell MR, Pittman GS, Faulkner EC, Watson MA, Bell DA. Expressionbased discovery of variation in the human glutathione S-transferase M3 promoter and functional analysis in a glioma cell line using allele-specific chromatin

immunoprecipitation. Cancer Res. 2005;65:99–104.

25. Wang S, Yang J, Ding C, Li J, You L, Dai M, et al. Glutathione S-Transferase mu-3

predicts a better prognosis and inhibits malignant behavior and glycolysis in

pancreatic cancer. Front Oncol. 2020;10:1539.

26. Tan X, Wang Y, Han Y, Chang W, Su T, Hou J, et al. Genetic variation in the GSTM3

promoter confer risk and prognosis of renal cell carcinoma by reducing gene

expression. Br J Cancer. 2013;109:3105–15.

27. Sun Y, Wang Y, Yin Y, Chen X, Sun Z. GSTM3 reverses the resistance of hepatoma

cells to radiation by regulating the expression of cell cycle/apoptosis-related

molecules. Oncol Lett. 2014;8:1435–40.

28. Singh H, Sachan R, Devi S, Pandey SN, Mittal B. Association of GSTM1, GSTT1, and

GSTM3 gene polymorphisms and susceptibility to cervical cancer in a North

Indian population. Am J Obstet Gynecol. 2008;198:301–3.

29. Loktionov A, Watson MA, Gunter M, Stebbings WS, Speakman CT, Bingham SA.

Glutathione-S-transferase gene polymorphisms in colorectal cancer patients:

Interaction between GSTM1 and GSTM3 allele variants as a risk-modulating

factor. Carcinogenesis. 2001;22:1053–60.

30. Kesarwani P, Singh R, Mittal RD. Association of GSTM3 intron 6 variant with

cigarette smoking, tobacco chewing and alcohol as modifier factors for prostate

cancer risk. Arch Toxicol. 2009;83:351–6.

31. Meding S, Balluff B, Elsner M, Schone C, Rauser S, Nitsche U, et al. Tissue-based

proteomics reveals FXYD3, S100A11 and GSTM3 as novel markers for regional

lymph node metastasis in colon cancer. J Pathol. 2012;228:459–70.

32. Cheng SY, Chen NF, Wen ZH, Yao ZK, Tsui KH, Kuo HM, et al. Glutathione

S-transferase m3 is associated with glycolysis in intrinsic temozolomide-resistant

glioblastoma multiforme cells. Int J Mol Sci. 2021;22:7080.

33. Chen Z, Zhou Y, Song J, Zhang Z. HCKSAAP_UbSite: improved prediction of

human ubiquitination sites by exploiting amino acid pattern and properties.

Biochim Biophys Acta. 2013;1834:1461–7.

34. Kinner A, Wu W, Staudt C, Iliakis G. Gamma-H2AX in recognition and signaling of

DNA double-strand breaks in the context of chromatin. Nucleic Acids Res.

2008;36:5678–94.

35. Byun JK, Lee S, Kang GW, Lee YR, Park SY, Song IS, et al. Macropinocytosis is an

alternative pathway of cysteine acquisition and mitigates sorafenib-induced

ferroptosis in hepatocellular carcinoma. J Exp Clin Cancer Res. 2022;41:98.

36. Liu B, Jiang S, Li M, Xiong X, Zhu M, Li D, et al. Proteome-wide analysis of

USP14 substrates revealed its role in hepatosteatosis via stabilization of FASN.

Nat Commun. 2018;9:4770.

37. Wang Y, Jiang Y, Ding S, Li J, Song N, Ren Y, et al. Small molecule inhibitors reveal

allosteric regulation of USP14 via steric blockade. Cell Res. 2018;28:1186–94.

38. Fako VE, Wu X, Pflug B, Liu JY, Zhang JT. Repositioning proton pump inhibitors as

anticancer drugs by targeting the thioesterase domain of human fatty acid

synthase. J Med Chem. 2015;58:778–84.

39. Lee BH, Lu Y, Prado MA, Shi Y, Tian G, Sun S, et al. USP14 deubiquitinates

proteasome-bound substrates that are ubiquitinated at multiple sites. Nature.

2016;532:398–401.

40. Li P, Jiang M, Li K, Li H, Zhou Y, Xiao X, et al. Glutathione peroxidase 4-regulated

neutrophil ferroptosis induces systemic autoimmunity. Nat Immunol.

2021;22:1107–17.

41. Zheng J, Sato M, Mishima E, Sato H, Proneth B, Conrad M. Sorafenib fails to trigger

ferroptosis across a wide range of cancer cell lines. Cell Death Dis. 2021;12:698.

42. Chen Y, Zhao Y, Yang X, Ren X, Huang S, Gong S, et al. USP44 regulates

irradiation-induced DNA double-strand break repair and suppresses tumorigenesis in nasopharyngeal carcinoma. Nat Commun. 2022;13:501.

43. Xie Z, Li W, Ai J, Xie J, Zhang X. C2orf40 inhibits metastasis and regulates chemoresistance and radio-resistance of nasopharyngeal carcinoma cells by influencing

cell cycle and activating the PI3K/AKT/mTOR signaling pathway. J Transl Med.

2022;20:264.

44. Hlouschek J, Ritter V, Wirsdorfer F, Klein D, Jendrossek V, Matschke J. Targeting

SLC25A10 alleviates improved antioxidant capacity and associated radioresistance of cancer cells induced by chronic-cycling hypoxia. Cancer Lett.

2018;439:24–38.

45. Ma X, Yao M, Gao Y, Yue Y, Li Y, Zhang T, et al. Functional immune cell-derived

exosomes engineered for the trilogy of radiotherapy sensitization. Adv Sci.

2022;9:e2106031.

46. Yuan L, Li S, Chen Q, Xia T, Luo D, Li L, et al. EBV infection-induced GPX4

promotes chemoresistance and tumor progression in nasopharyngeal carcinoma.

Cell Death Differ. 2022;29:1513–27.

47. Chen T, Jinlin D, Wang F, Yuan Z, Xue J, Lu T, et al. GSTM3 deficiency impedes

DNA mismatch repair to promote gastric tumorigenesis via CAND1/NRF2-

KEAP1 signaling. Cancer Lett. 2022;538:215692.

48. Sharma A, Alswillah T, Kapoor I, Debjani P, Willard B, Summers MK, et al. USP14 is

a deubiquitinase for Ku70 and critical determinant of non-homologous end

joining repair in autophagy and PTEN-deficient cells. Nucleic Acids Res.

2020;48:736–47.

49. Tsai Y, Xia C, Sun Z. The inhibitory effect of 6-Gingerol on ubiquitin-specific

peptidase 14 enhances autophagy-dependent ferroptosis and anti-tumor in vivo

and in vitro. Front Pharmacol. 2020;11:598555.

50. Shi D, Wu X, Jian Y, Wang J, Huang C, Mo S, et al. USP14 promotes tryptophan

metabolism and immune suppression by stabilizing IDO1 in colorectal cancer.

Nat Commun. 2022;13:5644.

51. Wang X, Liu M, Chu Y, Liu Y, Cao X, Zhang H, et al. O-GlcNAcylation of ZEB1

facilitated mesenchymal pancreatic cancer cell ferroptosis. Int J Biol Sci.

2022;18:4135–50.

52. Ates G, Goldberg J, Currais A, Maher P. CMS121, a fatty acid synthase inhibitor,

protects against excess lipid peroxidation and inflammation and alleviates cognitive loss in a transgenic mouse model of Alzheimer’s disease. Redox Biol.

2020;36:101648.

53. Bartolacci C, Andreani C, Vale G, Berto S, Melegari M, Crouch AC, et al. Targeting

de novo lipogenesis and the Lands cycle induces ferroptosis in KRAS-mutant lung

cancer. Nat Commun. 2022;13:4327.

54. Sun H, Cai H, Xu C, Zhai H, Lux F, Xie Y, et al. AGuIX nanoparticles enhance

ionizing radiation-induced ferroptosis on tumor cells by targeting the NRF2-GPX4

signaling pathway. J Nanobiotechnol. 2022;20:449.

55. Chen Q, Zheng W, Guan J, Liu H, Dan Y, Zhu L, et al. SOCS2-enhanced ubiquitination of SLC7A11 promotes ferroptosis and radiosensitization in hepatocellular

carcinoma. Cell Death Differ. 2022;30:137–151.

56. Wang Y, Zhao G, Condello S, Huang H, Cardenas H, Tanner EJ, et al. Frizzled-7

identifies platinum-tolerant ovarian cancer cells susceptible to ferroptosis. Cancer

Res. 2021;81:384–99.

57. Ding Y, Chen X, Liu C, Ge W, Wang Q, Hao X, et al. Identification of a small

molecule as inducer of ferroptosis and apoptosis through ubiquitination of GPX4

in triple negative breast cancer cells. J Hematol Oncol. 2021;14:19.

58. Zhou JC, Wu B, Zhang JJ, Zhang W. Lupeol triggers oxidative stress, ferroptosis,

apoptosis and restrains inflammation in nasopharyngeal carcinoma via AMPK/

NF-kappaB pathway. Immunopharmacol Immunotoxicol. 2022;44:621–31.

59. Xu X, Li Y, Wu Y, Wang M, Lu Y, Fang Z, et al. Increased ATF2 expression predicts

poor prognosis and inhibits sorafenib-induced ferroptosis in gastric cancer.

Redox Biol. 2023;59:102564.

60. Sun J, Zhou C, Zhao Y, Zhang X, Chen W, Zhou Q, et al. Quiescin sulfhydryl

oxidase 1 promotes sorafenib-induced ferroptosis in hepatocellular carcinoma by

driving EGFR endosomal trafficking and inhibiting NRF2 activation. Redox Biol.

2021;41:101942.

ACKNOWLEDGEMENTS

We thank Professor Mingyuan Chen (Sun Yat-sen University Cancer Center, China)

and Professor Rui You (Sun Yat-sen University Cancer Center, China) for providing

paraffin-embedded specimens of NPC patients. We also thank Dr. Xiuhua Yin (Suzhou

University, China) and Dr. Shengtang Liu (Suzhou University, China) for technical

support of protein-binding model.

AUTHOR CONTRIBUTIONS

YC and LC conceived the study and analysed the results. YC wrote the manuscript. YC

and YF designed and performed the experiments. YL, XZ, and LW were involved in

analyses and discussions of the data. YZ and KL collected the tumour specimens and

clinical data. LC and YL revised the manuscript and provided scientific directions. All

authors read and approved the final manuscript.

Y. Chen et al.

13

British Journal of Cancer

FUNDING

This work was supported by the National Natural Science Foundation of China

(82372970, 82102926), the Guangdong Basic and Applied Basic Research Foundation

(2023A1515010492, 2022A1515010083), the Guangzhou Basic and Applied Basic

Research Foundation (202201011049), the China Postdoctoral Science Foundation

(2022M711511), the Guangzhou Municipal Science and Technology Project (2022--

RC1348), and the President Foundation of Nanfang Hospital, Southern Medical

University (2021A006).

COMPETING INTERESTS

The authors declare no competing interests.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

All animal experiments were approved by the Institutional Animal Care and Use

Committee of Nanfang Hospital. All human tumour specimens were provided by the

Sun Yat‐sen University Cancer Center, Guangzhou, China. The informed consents

were obtained from all the participants prior to participation.

CONSENT FOR PUBLICATION

Not applicable.

ADDITIONAL INFORMATION

Supplementary information The online version contains supplementary material

available at https://doi.org/10.1038/s41416-024-02574-1.

Correspondence and requests for materials should be addressed to Longmei Cai.

Reprints and permission information is available at http://www.nature.com/

reprints

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims

in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons

Attribution 4.0 International License, which permits use, sharing,

adaptation, distribution and reproduction in any medium or format, as long as you give

appropriate credit to the original author(s) and the source, provide a link to the Creative

Commons licence, and indicate if changes were made. The images or other third party

material in this article are included in the article’s Creative Commons licence, unless

indicated otherwise in a credit line to the material. If material is not included in the

article’s Creative Commons licence and your intended use is not permitted by statutory

regulation or exceeds the permitted use, you will need to obtain permission directly

from the copyright holder. To view a copy of this licence, visit http://

creativecommons.org/licenses/by/4.0/.

© The Author(s) 2024

Y. Chen et al.

14

British Journal of Cancer