Pseudolaric acid B triggers ferroptosis in glioma cells via activation of Nox4 and inhibition of xCT

Accepted Manuscript Pseudolaric acid B triggers ferroptosis in glioma cells via activation of Nox4 and inhibition of xCT Zongqi Wang, Ye Ding, Xuanzhong Wang, Shan Lu, Chongcheng Wang, Chuan He, Lei Wang, Meihua Piao, Guangfan Chi, Yinan Luo, Pengfei Ge, M.D, Ph.D PII:

S0304-3835(18)30288-X

DOI:

10.1016/j.canlet.2018.04.021

Reference:

CAN 13861

To appear in:

Cancer Letters

Received Date: 15 March 2018 Revised Date:

16 April 2018

Accepted Date: 18 April 2018

Please cite this article as: Z. Wang, Y. Ding, X. Wang, S. Lu, C. Wang, C. He, L. Wang, M. Piao, G. Chi, Y. Luo, P. Ge, Pseudolaric acid B triggers ferroptosis in glioma cells via activation of Nox4 and inhibition of xCT, Cancer Letters (2018), doi: 10.1016/j.canlet.2018.04.021. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Abstract Ferroptosis is a form of programmed cell death decided by iron-dependent lipid peroxidation, but its role in glioma cell death remains unclear. In this study, we found Pseudolaric acid B (PAB) inhibited the viabilities of glioma cells in vitro and in vivo, which was accompanied by abnormal increases of intracellular ferrous iron, H2O2 and lipid peroxidation, as well as depletion of GSH

RI PT

and cysteine. In vitro studies revealed that the lipid peroxidation and the cell death caused by PAB were both inhibited by iron chelator deferoxamine, but exacerbated by supplement of ferric ammonium citrate. Inhibition of lipid peroxidation with ferrostatin-1 or GSH rescued

SC

PAB-induced cell death. Morphologically, the cells treated with PAB presented intact membrane, shrunken mitochondria with increased membrane density, and normal-sized nucleus without

M AN U

chromatin condensation. Mechanistically, PAB improved intracellular iron by upregulation of transferrin receptor. The increased iron activated Nox4, which resulted in overproduction of H2O2 and lipid peroxides. Moreover, PAB depleted intracellular GSH via p53-mediated xCT pathway, which further exacerbated accumulation of H2O2 and lipid peroxides. Thus, PAB triggers

AC C

EP

TE D

ferroptosis in glioma cells and is a potential medicine for glioma treatment.

ACCEPTED MANUSCRIPT Revised Manuscript

Original Research Article

Pseudolaric acid B triggers ferroptosis in glioma cells via activation of Nox4 and inhibition of xCT

RI PT

Zongqi Wang1,2, Ye Ding1,2, Xuanzhong Wang1,2, Shan Lu1,2, Chongcheng Wang1,2, Chuan He1,2, Lei Wang1,2, Meihua Piao3, Guangfan Chi4, Yinan Luo1,2, Pengfei Ge1,2*

SC

1 Department of Neurosurgery, first hospital of Jilin University, Changchun 130021, China. 2 Research center of neuroscience, first hospital of Jilin University, Changchun 130021, China. 3 Department of Anesthesiology, first hospital of Jilin University, Changchun 130021, China. 4Key Laboratory of Pathobiology, Ministry of Education, Jilin University, Changchun 130021, China

M AN U

* :corresponding author

Running Title: Pseudolaric acid B triggers ferroptosis in glioma cells

AC C

Words: 5500. Figures: 8. References: 53.

EP

TE D

Corresponding author: Pengfei Ge, M.D, Ph.D E-mail: [email protected], or [email protected] Address: Department of Neurosurgery, First Hospital, Jilin University. 71 Xinmin Avenue, Changchun 130021, Jilin Province, P.R.China.

ACCEPTED MANUSCRIPT Abstract Ferroptosis is a form of programmed cell death decided by iron-dependent lipid peroxidation, but its role in glioma cell death remains unclear. In this study, we found Pseudolaric acid B (PAB) inhibited the viabilities of glioma cells in vitro and in vivo, which was accompanied by abnormal

RI PT

increases of intracellular ferrous iron, H2O2 and lipid peroxidation, as well as depletion of GSH and cysteine. In vitro studies revealed that the lipid peroxidation and the cell death caused by PAB were both inhibited by iron chelator deferoxamine, but exacerbated by supplement of ferric ammonium citrate. Inhibition of lipid peroxidation with ferrostatin-1 or GSH rescued

SC

PAB-induced cell death. Morphologically, the cells treated with PAB presented intact membrane, shrunken mitochondria with increased membrane density, and normal-sized nucleus without

M AN U

chromatin condensation. Mechanistically, PAB improved intracellular iron by upregulation of transferrin receptor. The increased iron activated Nox4, which resulted in overproduction of H2O2 and lipid peroxides. Moreover, PAB depleted intracellular GSH via p53-mediated xCT pathway, which further exacerbated accumulation of H2O2 and lipid peroxides. Thus, PAB triggers

TE D

ferroptosis in glioma cells and is a potential medicine for glioma treatment.

AC C

EP

Key Words: Pseudolaric acid B; Ferroptosis; Nox4; xCT; Glioma

ACCEPTED MANUSCRIPT 1. Introduction Malignant glioma is the most common and aggressive form of intrinsic brain tumor [1]. Its therapy generally consists of surgical resection followed by radiotherapy in conjunction with

RI PT

alkylating agent temozolomide. However, the prognosis is still very poor, with a median survival of about 14.6 months [2]. Thus, novel therapeutic strategy is needed urgently. Iron plays a key role in the metabolism of all cells and is also found to regulate DNA replication, cell proliferation and

SC

oxidative stress [3]. Cancer cells exhibit an enhanced dependence on iron relative to their normal

M AN U

counterparts, which is termed iron addiction [3]. Although chelation of iron could sensitize glioma cells to temozolomide or radiotherapy [4, 5], the role of excessive intracellular iron in regulation of glioma cell destiny remains elusive.

Ferroptosis is a newly-established form of regulated cell death, which is coined after the

TE D

requirement for free ferrous iron [6]. Biochemically, the process of ferroptosis is characterized by accumulation of lipid peroxidation products and lethal reactive oxygen species (ROS) derived from iron metabolism[7]. Morphologically, the cells undergoing ferroptosis present normal-sized

EP

nuclei free of chromatin condensation and shrinking mitochondria with increased membrane

AC C

density [7].Ferroptosis is involved in multiple pathological processes including acute renal failure, ischemia/reperfusion injury and neurodegeneration [6]. It can be triggered in cancer cells by depleting glutathione (GSH) and amino acid cysteine or by inhibiting the phospholipid hydroperoxidase glutathione peroxidase 4 (GPX4) [8, 9], but is characteristically prevented by iron chelator deferoxamine and lipid peroxidation inhibitors ferrostatin, liproxstatin and zileuton[10]. Some chemical compounds, sorafenib, artesunate and piperlongumine, have been proven to kill cancer cells via activation of ferroptosis pathway [11-13]. Moreover, induction of

ACCEPTED MANUSCRIPT ferroptosis might overcome the resistance of cancer cells to current treatments, because ferroptosis-inducer erastin was reported to sensitize both acute myeloid leukemia cells, glioma cells and neck cancer cells to chemotherapeutic agents [14-16]. Therefore, triggering ferroptosis is

RI PT

emerging to be an effective approach to eliminate cancer cells. Pseudolaric acid B (PAB), a diterpene acid isolated from the root and trunk bark of Cortex pseudolaricis, has various bioactivities such as antimicrobial activity, antiangiogenic activity and

SC

antivirus activity [17,18]. Several reports have shown that PAB treatment could lead to death in

M AN U

human cancer cells from prostate cancer, cervical cancer and breast cancer [19-21]. Additionally, PAB was also found to inhibit cancer cell metastasis, circumvent cancer cell multidrug-resistance, and enhance cancer cell sensitivity to radiotherapy [22-25]. Therefore, PAB is a potential medicine in cancer treatment. Although previous study has shown that PAB could induce apoptosis and

TE D

autophagic death in cancer cells [19], accumulating evidences have demonstrated that ferroptosis could occur simultaneously with apoptosis and autophagy [26, 27]. In this study, we thus used rat and human glioma cell lines and mice model of xenograft glioma to investigate whether PAB

EP

could induce ferroptosis and its underlying mechanism.

AC C

2.Materials and Methods 2.1. Materials

Pseudolaric acid B, glutathione, ferric ammonium citrate and Vitamin E were purchased

from Sigma Company (St.Louis, MO). GKT137831 and ferrostatin-1were from selleckchem Company (Houston, TX). Deferoxamine, anti-p53(ab26), anti-Nox4(ab133303), anti-ferritin light chain(ab69090), anti-ferritin heavy chain(ab75972), anti-ferroportin(FPN, ab78066), anti-ferritin (ab82411), anti-transferrin receptor(ab1086), anti-GPX4(ab125066), and anti-xCT(ab175186)

ACCEPTED MANUSCRIPT were purchased from Abcam company (Cambridge, MA). Anti-phospho-p53(ser 15)(#9284) was from cell signaling technology company (Danvers, MA). Anti-β-actin antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). Other reagents were purchased from Sigma company (St.

RI PT

Louis, MO). 2.2. Cell lines and viability assay

Rat C6, human SHG-44, U87 and U251 glioma cells were obtained from Shanghai Institute of

SC

Cell Biology, Chinese Academy of Sciences (Shanghai, China). These cells were cultured in

M AN U

DMEM containing 10% fetal bovine serum, 2 mmol/L glutamine, penicillin (100 U/mL) and streptomycin (100µg/mL). The cells were maintained at 37 ºC and 5% CO2 in a humid environment and the cells in the mid-log phase were used. Cellular viability was assessed using an MTT assay and was expressed as a ratio to the absorbance value at 570 nm of the control cells.

TE D

2.3. Lactate dehydrogenase release cell death assay

Cell death rate was decided by using lactate dehydrogenase cytotoxicity assay kit (Beyotime Biotech, Nanjing, China). As describe by the manufacturer’s instructions, the absorbance value of

EP

each sample was read at 490nm. The death ratio was calculated using the following formula: cell

AC C

death ratio %= (A sample – A control / A max – A control) × 100. A sample: sample absorbance value; A control: the absorbance value of control group; A max: the absorbance value of positive group

2.4. Transmission electron microscopy The collected U251 cells by centrifugation were fixed in ice-cold 2.5% glutaraldehyde in PBS (pH 7.3), rinsed with PBS and post-fixed in 1% osmium tetroxide with 0.1% potassium ferricyanide, dehydrated through a graded series of ethanol (30–90%) and embedded in Epon

ACCEPTED MANUSCRIPT (Energy Beam Sciences, Agawam, MA, USA). Semithin (300 nm) sections were cut using a Reichart Ultracut, stained with 0.5%toluidine blue and examined under a light microscope. Ultrathin sections (65 nm) were stained with 1% uranyl acetate and 0.1% lead citrate, and

RI PT

examined on a JEM2000EX transmission electron microscope (JEOL, Pleasanton, CA, USA). 2.5. Iron assay

Intracellular ferrous iron level was determined using an iron colorimetric assay kit purchased

SC

from Biovision (Milpitas, California, USA). According to the manufacturer's instructions, the

M AN U

collected cells and 10mg glioma tissues were respectively added into iron assay buffer, homogenized on ice, and centrifuged at 16000×g for 10 min at 4˚C to obtain the supernatant for assay. 50µL sample was incubated with 50µL assay buffer in a 96-well microplate for 30min at 25˚C. Then, the sample was incubated with100µL iron probe in the dark for 60min at 25˚C, and

TE D

measured with a microplate reader at a wavelength of 593 nm. Absorbance values were calibrated to a standard concentration curve to calculate the concentration of iron. The results were expressed as a ratio to the concentration of the control cells.

EP

2.6. Lipid peroxidation assessment

AC C

Lipid peroxidation was assessed by using a malondialdehyde (MDA) assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer protocol. Briefly, the collected cells and 10mg glioma tissue were respectively added into lysis buffer, homogenized on ice, and centrifuged at 1600×g for 10min at 4˚C to obtain the supernatant. After the protein concentration was measured, 100µL sample was incubated with 100µL test solution for 15min at 100˚C. When cooling down to room temperature, the samples were centrifuged at 1000×g for 10min to get the supernatant, which was read at 530 nm in a microplate reader. MDA

ACCEPTED MANUSCRIPT content was expressed as a ratio to the absorbance value of the control cells. 2.7. NADPH oxidase activity assay The activity of NADPH oxidase was analyzed by using a lumiometric assay kit from Genmed

RI PT

Scientifics Inc. U.S.A (Arlington, MA) according to the manufacturer’s instruction. Briefly, the cells were scraped off in reagent A and centrifuged at 300×g for 5 min at 4˚C. The pelleted cells were suspended in reagent B and homogenized on ice. The homogenates were votex for 15s and

SC

placed on ice for 30min, centrifuged at 16000×g for 5min at 4˚C to obtain the supernatant. The

M AN U

pellet was suspended in reagent B and further homogenized. After the protein concentration was measured, 100µl samples containing 50µl supernatant and 50µl pellet was incubated the test solution (a mixture of reagent C, reagent D and reagent F) for 2min at 30˚C, and measured immediately with a luminometer to get relative light unit. Finally, the results were expressed as a

TE D

ratio to the relative light unit of the control cells.

2.8. Measurement of total intracellular glutathione (GSH) Intracellular total GSH was measured by using the DTNB-GSSH reductase recycling assay

EP

kit (Beyotime Biotechnology, Nanjing, China) as described by manufacture. Briefly, the collected

AC C

cells and 10mg glioma tissues were respectively added into protein-removing buffer S, homogenized on ice with a homogenizer, and centrifuged at 10,000×g for 10 min at 4˚C to get the supernatant used for intracellular total GSH assay. GSH content was expressed as a ratio to the absorbance value at 412 nm of the control cells. 2.9. Measurement of intracellular cysteine Intracellular cysteine was measured by using a cysteine assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer protocol. Briefly, the

ACCEPTED MANUSCRIPT collected cells and 10mg xenograft glioma tissues were respectively added into reagent A, homogenized on ice, and centrifuged at 8000×g for 4min at 4˚C to obtain the supernatant for assay. After the protein concentration was measured, 20µL sample was incubated with 100µL reagent B

RI PT

and 100µL reagent C for 15min at room temperature and read at absorbance 600 nm in a microplate reader. Finally, the results were expressed as a ratio to the absorbance value of the control cells.

SC

2.10. Intracellular H2O2 content assay

M AN U

The content of intracellular H2O2 was analyzed with a H2O2 assay kit (Beyotime Biotech, Nanjing, China) according to the manufacturer protocol. Briefly, the collected cells and 10mg glioma tissues were respectively added into lysis buffer, homogenized on ice, and centrifuged at 12000×g for 5 min at 4˚C to obtain the supernatant. Then, 50µl of supernatant and 100µl of test

TE D

solution were added into a tube, placed at room temperature for 30 min, and measured with a microplate reader at a wavelength of 560 nm. Absorbance values were calibrated to a standard concentration curve to calculate the concentration of H2O2. Finally, the results were expressed as a

EP

ratio to the concentration of the control cells.

AC C

2.11. Detection of superoxide anion The average level of superoxide anion was detected by dihydroethidium (DHE, Beyotime

Biotech, Nanjing, China). The cells were washed twice in PBS and then loaded with DHE (10 µmol/L) in fresh medium at 37 °C in the dark for 30 min. The fluorescence was measured at an excitation wavelength of 485 nm and an emission wavelength 530 nm. The levels of superoxide anions were expressed as a ratio to the absorbance value of control cells. Other groups of C6 and SHG-44 cells stained with DHE as described above were use to observe under fluorescence

ACCEPTED MANUSCRIPT microscope (Olympus IX71, Tokyo, Japan). 2.12. Mitochondrial membrane potential assay Mitochondrial membrane potential was assayed using JC-1 according to manufacture’s

RI PT

instruction (Beyotime Biotech, Nanjing, China). The percent of cells with a loss in mitochondrial membrane potential was determined by gating the population of cells with a decrease in JC-1 aggregates but increase in JC-1 monomers. After being stained with JC-1, the ells were analyzed

SC

by flow cytometry (FACScan, Becton Dickinson, San Jose, CA). The excitation wavelength of

M AN U

JC-1 is 488 nm, and the approximate emission wavelength of the monometric and J-aggregate forms is 529 and 590 nm, respectively. Additionally, one group of cells stained with JC-1 were observed under fluorescence microscope (Olympus IX71, Tokyo, Japan), and the other group of cells stained with JC-1 were used to measure the intensity of red (excitation wavelength: 530nm;

wavelength: 528).

TE D

emission wavelength: 590) and green fluorescence (excitation wavelength: 485nm; emission

2.13. Transfection of small interfering RNA (SiRNA)

EP

Transfection of SiRNA was performed by using Lipofectamine 2000 (Invitrogen, USA)

AC C

according to the manufacturer’s instructions. The NOX4 siRNA (5′-CCUCAGCAUCUGUUCUUAATT-3′), GPX4 siRNA (5′-UGGUGAUAGAGAAGGACCUTT-3′) and scrambled siRNA (5′-UUCUCCGAACGUGUCACGUTT-3′) were purchased from GenePharma Company (Suzhou, China). 2.14. Rat C6 tumor xenograft in mice Twenty athymic BALB/c nude mice (aged 4 weeks, weight 20-22 g, from Shanghai

ACCEPTED MANUSCRIPT laboratory animal Center, Shanghai, China) were housed in a specific pathogen-free environment under the condition of 12-h light/12-h dark cycle, free access to food and water, and acclimatised to their surroundings for three days. The study was approved by the ethics committee of the first

RI PT

hospital of Jilin University (Changchun, China). A total of 1 × 106 logarithmically growing C6 cells in 100 µl of PBS were subcutaneously injected into the right flank of each mouse. Therapeutic experiments were started when the tumour reached about 150 mm3 after about 7 days.

SC

The mice were allocated to receive intraperitoneal injections of vehicle (control group, n

M AN U

=5/group), PAB at the dosage of 10mg/kg body weight (n =10/group) and 20mg/kg body weight (n =10/group) in the same volume 50 µL once a days for 8 times. The tumor size was measured using a slide caliper, and the tumor volume was calculated using the formula: 0.5 ×A× B2, in which A is the length of the tumor and B is the width. On the next day of the last treatment, the

TE D

mice were euthanized by cervical dislocation and tumor tissues were excised. The tumor tissues were frozen immediately in liquid nitrogen for western blotting analysis. 2.15. Gel Electrophoresis and western Blotting

EP

The cell collection and homogenization were performed as described previously [28]. Then,

AC C

the homogenates were centrifuged at 1000 g for 10 min at 4 ºC to obtain the supernatants, the protein content of which was determined using Bio-Rad protein assay kit. After SDS electrophoresis and transfer, the PVDF membranes were blocked with 3% bovine serum albumin in TBS for 30 min at room temperature, then incubated overnight at 4 ºC with primary antibodies. After being incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (1:2000), blots were washed and immunoreactive proteins were visualized on a Kodak X-omat LS film (Eastman Kodak Company, New Haven, CT) with an enhanced chemiluminescence (Amersham Biosciences,

ACCEPTED MANUSCRIPT Piscataway NJ). Densitometry was performed with Kodak ID image analyses software. 2.16. Statistical analysis All data represent at least 4 independent experiments and are expressed as mean±SD.

considered to represent statistical significance. 3. Results

RI PT

Statistical comparisons were made using One-way ANOVA. P-values of less than 0.05 were

SC

3.1. Pseudolaric acid B inhibited viability and induced death in glioma cells

M AN U

To investigate the toxic effect of PAB on glioma cells, we examined PAB-induced changes in the viabilities of C6, SHG-44, U251 and U87 glioma cells by using MTT assay. As shown in Fig1.B-E, in comparison with control cells, the viabilities of the C6, SHG-44, U251 and U87 cells were decreased drastically by PAB in a concentration- and incubation time- dependent manner.

TE D

Moreover, LDH release assay proved that PAB at indicated concentration induced glioma cell death at incubation 24h, and the cell death was dependent on the PAB concentration (Fig.1F). Therefore, these data suggested that PAB not only inhibited the viabilities of glioma cells, but also

EP

triggered glioma cell death. Then, we used cell line HL-7720 (hepatic immortal cell line) to

AC C

evaluate the toxicity of PAB in normal cells. MTT assay showed the viabilities of HL-7702 cells did not reduced markedly after 48h incubation with PAB at indicated dosage (Fig. 1G). Thus, PAB had limited toxic effect on normal cells. On the basis of the reduction caused by PAB in the cellular viability of the C6, SHG-44, U87 and U251 cells, we calculated the IC50 values of PAB at 48h. They were 2.35µmol/L, 2.30µmol/L, 2.17µmol/L, and 2.49µmol/L for the C6, SHG-44, U251 and U87 cells, respectively. Thus, 2µmol/L was used as the IC50 value of PAB in the subsequent studies.

ACCEPTED MANUSCRIPT 3.2. Iron regulated PAB-induced glioma cell death To test whether PAB induced ferroptosis in glioma cells, we assayed PAB-induced changes in intracellular ferrous irons. When compared with that in the control cells, ferrous iron was

RI PT

improved apparently by 0.5µmol/L PAB at incubation 12h in C6, SHG-44, U251 and U87 cells, and further increased when PAB concentration was elevated to 2.0µmol/L or the incubation time

time- and dosage-dependent manner (Fig. 3A).

SC

was extended to 24h. These indicated that intracellular ferrous iron was upregulated by PAB in a

M AN U

For elucidating the significance of iron increase, the cells were treated with iron chelator deferoxamine (DFO) at 500µmol/L for 1h and then incubated with 2.0µmol/L PAB for 24h. We found that DFO markedly inhibited PAB-induced increase of intracellular ferrous iron (Fig.3B). Moreover, LDH release assay showed that the glioma cell death caused by PAB was apparently

TE D

inhibited in the presence of DFO (Fig.3C). Consistently, observation under light microscope proved that DFO pretreatment obviously reversed PAB-induced reduction in cell size (Fig.2A). In contrast, pretreatment with 500µmol/L ferric ammonium citrate (FAC) for 1h exacerbated the

EP

lethal effect of PAB on glioma cells (Fig.3D). These indicated that PAB triggered glioma cell

AC C

death via improvement of intracellular iron. Furthermore, transmission electronic microscopy revealed that the U251 cells treated with

2.0µmol/L PAB for 12h had intact membrane with decreased microvilli, shrunken mitochondria with increased membrane density, and normal-sized nucleus free of chromatin condensation or margination, when compared with the control cells (Fig.2 B). Then, we examined mitochondrial membrane potential by using probe JC-1 which aggregates in mitochondria and present high red fluorescence in normal cells, whereas exists in the cytoplasm as a monomer emitting green

ACCEPTED MANUSCRIPT fluorescence in the cells to die. Both fluorescence microscopy and statistical analysis of the fluorescence intensity showed that PAB treatment resulted in reduction in the red fluorescence, but elevation in the green fluorescence at incubation 24(Fig.2 C and D). Moreover, flow cytometry

RI PT

analysis demonstrated as well that PAB treatment resulted in reduction of mitochondrial membrane potential (Fig.2 E). These data indicated that that PAB induced mitochondrial depolarization. These were consistent with the morphological characteristics of ferroptosis [10,

SC

29].

M AN U

To unveil the regulators responsible for PAB-induced abnormal increase of intracellular iron, western blotting analysis was performed. Transferrin (TF) that could bind with iron and transferrin receptor (TFR) that accounts for transferring transferrin-iron complex into cells were both time-dependently upregulated by 2.0µmol/L PAB in all the four glioma cell lines, despite ferritin

TE D

(FTH, FTL) and ferroportin (FPN) which could mitigate intracellular ferrous iron were also upregulated (Fig.3 E). This indicated that PAB induced accumulation of iron in glioma cells via promotion of iron importation, not through inhibition of iron exportation.

EP

3.3. Iron contributed to PAB-induced glioma cell death via causing lipid peroxidation

AC C

Because lipid peroxidation is a feature of ferroptosis, we examined PAB-induced changes in lipid peroxide. We found that MDA, a final product of lipid peroxidation, was increased significantly by 0.5µmol/L PAB at incubation 12h, and aggravated when incubation time was extended to 24h or PAB dosage was elevated to 2µmol/L (Fig.4 A). This indicated that PAB induced lipid peroxidation in a time- and dosage-dependent manner. In contrast, the generation of MDA induced by PBA was markedly inhibited in the presence of DFO, but exacerbated when the cells was supplemented with FAC (Fig.4 B). This indicated that iron regulated PLA-induced lipid

ACCEPTED MANUSCRIPT peroxidation. To investigate the role of lipid peroxidation in PAB-induced glioma cell death, the C6 and SHG-44 cells were pretreated with ferrostatin-1(Fer-1) that could specifically inhibit oxidative

RI PT

lipid damage [30]. We found that pretreatment with 50µmol/L Fer-1 for 1h markedly prevented PAB-induced overproduction of MDA (Fig.4 B). Moreover, LDH release assay proved that Fer-1 pretreatment attenuated the glioma cell death caused by PAB (Fig.4 C), which was also confirmed

SC

by light microscopy (Fig.2 A). Consistently, pretreatment with 100µmol/L Vitamin E or 10mmol/L

M AN U

GSH for 1h not only inhibited the level of MDA (Fig.4 B), but also mitigated the lethal effect of PAB on glioma cells (Fig.4 C). Therefore, these results suggested that lipid peroxidation contributed to PBA-induced glioma cell death.

3.4. Iron modulated PAB-induced overproduction of H2O2 and superoxide

TE D

Given that lipid peroxidation could be initiated by hydroxyl radicals that are released from hydrogen peroxide by ferrous iron via Fenton type reaction [31], we compared the differences in the levels of intracellular H2O2 between the cells treated with and without PAB. As shown in

EP

Fig.4D, compared with that in the control cells, H2O2 accumulated in the cell treated with

AC C

0.5µmol/L PAB for 12h, which became more obvious when PAB dosage was elevated to 2.0µmol/L or the incubation time was extended to 24h. Considering that H2O2 could be converted from superoxide [32], we investigated the effect of

PAB treatment on superoxide generation. Fluorescence microscopy revealed that the intensity of the red fluorescence detected by superoxide probe DHE was obviously stronger in the cells treated with 2.0µmol/L PAB for 24h, when compared with that in control cells (Fig. 4E). Statistical analysis of the fluorescence intensity proved that incubation with 0.5µmol/L PAB for 12h could

ACCEPTED MANUSCRIPT lead to excessive generation of superoxide, and the levels of superoxide were positively dependent on the PAB concentration or the incubation time (Fig. 4F). Thus, these data indicated that PAB induced overproduction of superoxide in glioma cells.

RI PT

However, PAB-induced overproduction of superoxide and H2O2 were both significantly prevented by DFO pretreatment (Fig.4 G), but promoted by supplement with FAC (Fig.4 H). This indicated that iron regulated PAB-induced overproduction of H2O2 and superoxide.

SC

3.5. Iron contributed to PAB-induced activation of NADPH oxidase 4

M AN U

Because superoxide could be produced by activated NADPH oxidase and glioma cells have higher levels of Nox4[33], we speculated that PAB might over-activate Nox4 in glioma cells. As shown in Fig.5A, the activity of Nox4 was increased markedly by 0.5µmol/L PAB at incubation 12h, and further improved when PAB dosage was elevated to 2.0µmol/L or the incubation time

TE D

was extended to 24h. Then, western blotting revealed that Nox4 was upregulated by PAB in a time-dependent manner in the C6, SHG-44, U251 and U87 cells (Fig. 5B). Notably, the increased activity and the upregulated expression of Nox4 caused by PAB were both inhibited in the

EP

presence of DFO, but reinforced by FAC (Fig. 5C and D). This indicated that iron regulated the

AC C

activity and expression of Nox4.

To verify the role of Nox4 in PAB-induced generation of H2O2 and lipid peroxides, the cells

were pretreated with Nox4 inhibitor GKT137831 at 500µmol/L for 1h and then incubated with 2.0µmol/L PAB for 24h. It was found that GKT137831 not only inhibited PAB-induced increase in the activity of NADPH oxidase (Fig. 5B), but also prevented PAB-induced overproduction of superoxide (Fig.4G), hydrogen peroxide (Fig.4H), and MDA (Fig.4B). Therefore, these data indicated that Nox4 contributed to PAB-induced overproduction of H2O2 and lipid peroxides in

ACCEPTED MANUSCRIPT glioma cells. Additionally, LDH release assay revealed that GKT137831 prevented PAB-induced death in the SHG-44 and U251 cells (Fig.5E), which was confirmed as well by the observation under light

RI PT

microscope (Fig.2A). Because the upregulated expression of Nox4 caused by PAB was obviously inhibited by GKT137831 (Fig.5F), we introduced SiRNA to knock down Nox4 and examined its effect on PAB-induced toxicity in glioma cells. We found that knockdown of Nox4 with SiRNA

SC

made the cells resistant to PAB-induced death (Fig.5G and H), as well as inhibited the increases of

M AN U

intracellular H2O2 and MDA caused by PAB (Fig.5 I and J). Therefore, these data indicated that activated Nox4 contributed to PAB-induced overproduction of intracellular H2O2 and lipid peroxide and glioma cell death.

3.6. GSH depletion contributed to PAB-induced accumulation of H2O2 and lipid peroxides

TE D

Because GSH could be used by GPX4 to repair oxidative damage in lipids and reduce intracellular H2O2 [34], we examined PAB-induced changes in GPX4 and GSH. As revealed by western blotting, the protein level of GPX4 was upregulated time-dependently by 2.0µmol/L PAB

EP

in the C6, SHG-44, U87 and U251 cells (Fig.6A). Knockdown of GPX4 with SiRNA not only

AC C

prevented its upregulation caused by PAB, but also aggravated PAB-induced glioma cell death (Fig.6B and C). This indicated that GPX4 inhibited PAB-induced glioma cell death. However, both GSH and cysteine that is converted from cystine and then used as a material

used for GSH synthesis decreased significantly by 0.5µmol/L PAB at incubation 12h, and the decreases were exacerbated when PAB concentration was increased or the incubation time was extended (Fig6. D and E). In contrast, supplement of 10mmol/L GSH not only obviously mitigated PAB-induced improvement of intracellular H2O2 and MDA, but also inhibited glioma cell death

ACCEPTED MANUSCRIPT (Fig.4H, B and C), indicating that GSH depletion contributed to PAB-induced accumulation of H2O2 and lipid peroxide, and glioma cell death. Then, xCT(SLC7A11) that is a specific light-chain subunit of the cystine/glutamate antiporter

RI PT

accounting for transporting extracellular cystine into cells and p53 which could inhibit xCT transcription were analyzed by using western blotting[35] . We found xCT was downregulated by 2.0µmol/L PAB in a time-dependent manner, whereas p53 and phospho-p53 were both

SC

upregulated concomitantly in C6, SHG-44, U251 and U87 cells (Fig6.A). Therefore, these results

M AN U

indicated that PAB induced GSH depletion via activation of p53, which resulted in xCT inhibition. Notably, we also found that knockdown of GPX4 with siRNA improved PAB-induced overexpression of TFR and phosphorylation of p53, but exacerbated the downregulation of xCT (Fig.6B). This further demonstrated that GPX4 inhibited PAB-induced glioma cell death.

TE D

3.7. PAB improved ferrous iron and lipid oxidation in vivo.

To further evaluate the lethal effect of PAB on glioma cell, nude mice with subcutaneous xenograft glioma were treated with PAB. As shown in Fig.7A, in comparison with that in the

EP

control group, the volume of xenograft glioma became markedly smaller after eight times

AC C

treatment with PAB at the dosage of 10mg/kg weight each day. Moreover, the tumor volume decreased more apparently when the mice received treatment with PAB at the dosage of 20mg/kg body weight (Fig.7B). This indicated that PAB inhibited the growth of glioma cell in vivo. When the mice were sacrificed at day 9, the levels of ferrous iron in the removed glioma

tissue was assayed. We found iron was significantly increased by PAB at the dosage of 10mg/kg body weight, and reached a higher level when PAB dosage was increased to 20mg/kg body weight (Fig.7C). Western blotting analysis revealed that, the protein levels of TF and TFR were both

ACCEPTED MANUSCRIPT significantly increased by PAB, despite FTH, FTL, and FPN were all also upregulated (Fig.7D). This indicated that the increases of iron ions caused by PAB might be associated with the upregulation of TFR.

RI PT

Additionally, PAB treatment also resulted in dosage-dependent overproduction of MDA and H2O2 (Fig.7E and F), and depletion of GSH and cysteine (Fig.7G and H). Western blotting revealed that PAB upregulated the expression of Nox4, GPX4, p53 and phospho-p53, whereas

M AN U

H2O2 and MDA and depletion of GSH and cysteine.

SC

downregulated xCT (Fig.7D), which might explain why PAB could induce overproduction of

Therefore, the in vivo data suggested that the inhibitory effect of PAB on the growth of glioma cell in vivo was associated with increased intracellular iron and lipid peroxidation. 4. Discussion

TE D

In summary, we found in this study that PAB inhibited the viabilities of glioma cells in vitro and in vivo, which was accompanied by abnormal increases of intracellular ferrous iron, H2O2 and lipid peroxides, as well as depletion of GSH and cysteine. In vitro studies revealed that the lipid

EP

peroxidation and the cell death caused by PAB were both exacerbated by supplement of FAC, but

AC C

inhibited by DFO. Inhibition of lipid peroxidation with Fer-1 or GSH rescued PAB-induced glioma cell death. Morphologically, the cells treated with PAB presented intact membrane with decreased microvilli, shrunken mitochondria with increased membrane density, and normal-sized nucleus without chromatin condensation or margination. Furthermore, we proved that PAB improved intracellular iron by upregulation of transferrin receptor. PAB induced lipid peroxidation not only through activating Nox4 in an iron-dependent manner, but also by depleting GSH and cysteine via p53-regulated inhibition of xCT (Fig. 8).

ACCEPTED MANUSCRIPT Ferroptosis has been established as a non-apoptotic regulated cell death. Notably, it was found that malignant glioma cells potently extract iron from the microenvironment more effectively than other tumor cells [36], suggesting that induction of ferroptosis might be a reasonable strategy to

RI PT

eliminate glioma cells. Although ferroptosis could be induced via different pathways by different agents [8, 9], the key feature of ferroptosis is that the inducible death could be prevented not only by iron chelators, but also by lipid peroxidation inhibitors [8, 9]. Consistently, we found in this

SC

study that PAB induced abnormal increase of intracellular ferrous iron and lipid peroxidation

M AN U

during the process of glioma cell death. Chelation of intracellular iron with deferoxamine and inhibition of lipid peroxidation with Fer-1 or Vitamin E protected glioma cell against PAB-induced death. Additionally, transmission electronic microscopy revealed that PAB-induced morphological changes in glioma cells were also in line with the characteristics of ferroptosis [7], but different

cells.

TE D

with apoptosis, necroptosis and autophagy. Thus, we think that PAB induced ferroptosis in glioma

Abnormal increase of intracellular ferrous iron resulting from disrupted balance between

EP

iron uptake, storage and export is regarded as an initiating factor leading to ferroptosis [2].

AC C

Transferrin receptor (TFR) located at cell membrane is responsible for importing extracellular ferric iron into cells by internalizing the transferrin-iron complex through receptor-mediated endocytosis, but ferroportin (FPN) in the cell plasma membrane exports ferrous iron from cells[2]. If intracellular ferrous iron is not used immediately for cell process, it will be stored as ferric iron by ferritin which composed of subunits of the heavy chains (FTH) and light chains (FTL)[2]. Upregulation of transferrin receptor is a pathway to improve intracellular iron level and induce ferroptosis. It was reported that upregulation of transferrin receptor could improve intracellular

ACCEPTED MANUSCRIPT iron level and result in ferroptosis [37], but genetic silence of transferrin receptor prevented the occurrence of ferroptosis [38]. Consistently, we demonstrated in this study that PAB significantly upregulated the expression of transferrin receptor in a time-dependent manner and increased the

RI PT

level of transferrin, which might explain why PAB induced abnormal increase of ferrous in glioma cells. Downregulation of ferritin or ferroportin was also a pathway to increase intracellular iron and trigger ferroptosis. It was found that the ferroptotic cell death caused by combined treatment

SC

with siramesine and lapatinib in breast cancer cells was alleviated by overexpression of ferroportin,

M AN U

but reinforced when ferroportin was knocked down [39]. Similarly, knockdown of ferritin could also enhance the sensitivities of cancer cells to chemotherapeutic agents [40]. In this study, we found that FTH, FTL and FPN were all upregulated by PAB. Thus, PAB improved intracellular iron levels via upregulation of transferrin receptor, not via inhibition of ferritin or ferroportin.

TE D

Additionally, the protein levels of transferrin receptor, ferritin and ferroportin are also decided by intracellular iron levels. Iron deficiency could upregulate the expression of transferrin receptor to increase iron intake [41], whereas excessive irons would upregulate the expression of ferritin and

EP

ferroportin to export iron [42]. Thus, we think that the increase of FTH, FTL, and FPN caused by

AC C

PAB was an adaptive stress response to PAB-induced increase of intracellular iron. Lipid peroxidation caused by ferrous iron is thought to be a crucial step leading to

ferroptotic cell death [8]. Within cells, polyunsaturated fatty acids that exist in lipoproteins and lipid-containing structures are prone to being attacked by ROS, which leads to formation of lipid peroxides [43]. Among ROS, hydroxyl radical that could be released from H2O2 by ferrous iron via Fenton reaction is highly reactive and damaging [31, 32]. H2O2 is often converted from superoxide in the presence of superoxide dismutase, and NADPH oxidase is an enzyme

ACCEPTED MANUSCRIPT accounting for superoxide generation [33]. Nox4 is the first identified nonphagocytic NADPH oxidase. Glioma cells have higher expressional levels of Nox4, but lack of Nox1, Nox2, Nox3 and Nox5 [33]. Moreover, Nox4 could constitutively produce high level of H2O2 [44]. Although it was

RI PT

reported that PAB induced overproduction of ROS in murine fibrosarcoma L929 cells [45] and Nox1 was involved in regulation of erastin-induced ferroptosis [8], we found in this study that the protein level and the activity of Nox4 were both obviously upregulated by PAB in glioma cells.

SC

Inhibition of Nox4 with its specific inhibitor GKT137831 or knockdown of Nox4 with SiRNA

M AN U

significantly prevented PAB-induced excessive generation of H2O2 and lipid peroxide MDA, as well as glioma cell death. Consistent with previous report showing that the activity of NADPH oxidase was regulated by intracellular ferrous iron [46], the data in this study proved the increased activity of Nox4 resulting from PAB treatment was prevented by iron chelator DFO, but

TE D

reinforced by FAC. Moreover, we used nude mice model with xenograft glioma in this study to demonstrate that PAB induced significant increase of ferrous iron in glioma, which might explain why Nox4 could be activated by PAB in vivo. Thus, we think that iron contributed to

EP

PAB-induced lipid peroxidation via two pathways, one is to activate Nox4 to produce H2O2 and

AC C

the other is to react with H2O2 to produce hydroxyl radicals. GPX4/GSH plays a crucial role in protecting cells against the damage caused by oxidative

stress. GPX4 catalyzes the reduction of hydrogen peroxide, organic hydroperoxides, and lipid peroxides at the expense of GSH[34]. Pharmacological and genetic inhibition of GPX4 could result in ferroptosis, whereas high levels of GPX4 confered resistance to ferroptosis activation [9]. In glioma cells, knockdown of GPX4 not only inhibit the proliferation of glioma cells, but also induce glioma cell death [47]. Consistently, we found in this study that PAB upregulated the

ACCEPTED MANUSCRIPT protein levels of GPX4 and knockdown of GPX4 enhanced PAB-induced glioma cell death. Thus, PAB did not induce ferroptosis via inhibition of GPX4. Depletion of GSH via inhibition of sodium-independent cystine-glutamate antiporter xCT

RI PT

was also found to be an efficient pathway leading to ferroptosis in cancer cells [8]. xCT accounts for transportation of extracellular cystine into cells, and then cystine is reduced to cysteine used for GSH synthesis. Although higher level of xCT predicts poor survival in the patients with

SC

malignant glioma [48], inhibition of xCT could effectively sensitize gliomas to radiotherapy [49].

M AN U

In this study, we found that PAB downregulated the protein level of xCT and induced depletion of both cysteine and GSH in glioma cells, but supplement of GSH inhibited PAB-induced glioma cell death, as well as overproduction of ROS and lipid peroxide. Thus, inhibition of xCT is a pathway via which PAB triggered ferroptosis in glioma cells. Previous reports showed that ATF4 and tumor

TE D

suppressor p53 play opposite roles in regulation of the expression of xCT. Activated ATF4 upregulated xCT at transcriptional level in glioma cells [50], whereas activated p53 downregulated the protein levels of xCT [51]. Although we did not examine PAB-induced changes in ATF4, the

EP

data in this study showed that PAB activated p53 because the phospho-p53 was upregulated by

AC C

PAB in a time-dependent manner in C6 and SHG-44 glioma cells, even in the U87 cells with wild-type p53 and the U251 cells harboring mutant p53. Similarly, it was also reported that the lethal effects caused by PAB on human lung cancer cells and gastric carcinoma cells were also related to p53 activation [52, 53]. Therefore, these studies suggested that PAB inhibited xCT through activation of p53. In conclusion, we demonstrated in this study that ferroptosis is a pathway leading to PAB-induced glioma cell death. PAB improved intracellular iron via upregulation of transferrin

ACCEPTED MANUSCRIPT receptor. The increased intracellular ferrous iron activated Nox4, which resulted in overproduction of H2O2 and contributed to lipid peroxidation. Moreover, PAB depleted intracellular GSH and cysteine via p53-mediated xCT pathway, which further enhanced intracellular accumulation of

RI PT

H2O2 and lipid peroxidation. Acknowledgments

This work was supported by National Nature and Science Foundation of China (81372697,

SC

81772669), Changbaishan Scholar Project of Jilin Province (2013026), and Scientific Research

M AN U

Foundation of Jilin province (20160101127JC). We thank Dr. Chao Niu for his kind providing HL-7702 cell line for us. Conflict of Interests

The authors declare that they have no conflict of interests.

TE D

References

[1] M. Silginer, M. Weller, R. Stupp, P. Roth, Biological activity of tumor-treating fields in preclinical glioma models, Cell Death Dis 8(2017) e2753.

EP

[2] C. Legendre, E.Garcion, Iron metabolism: a double-edged sword in the resistance of glioblastoma to therapies, Trends Endocrinol Metab 26(2015) 322-331.

AC C

[3] D.H. Manz, N.L. Blanchette, B.T. Paul, F.M. Torti, S.V.Torti, Iron and cancer: recent insight, Ann N Y Acad Sci 1368(2016) 149-161. [4] G.A.Alexiou, P. Gerogianni, E. Vartholomatos, A.P.Kyritsis, Deferiprone enhances temozolomide cytotoxicity in glioma cells, Cancer Invest 34(2016) 489-495. [5] S.D.Ivanov, A.L. Semenov, E.G. Kovan'ko, V.A.Yamshanov, Effects of iron ions and iron chelation on the efficiency of experimental radiotherapy of animals with gliomas, Bull Exp Biol Med 158(2015) 800-803. [6] H.O. Fearnhead, P. Vandenabeele, T.Vanden Berghe, How do we fit ferroptosis in the family of regulated cell death?, Cell Death Differ 24(2017) 1991-1998.

ACCEPTED MANUSCRIPT [7] J.Y. Cao, S.J.Dixon, Mechanisms of ferroptosis, Cell Mol Life Sci 73(2016) 2195-2029. [8] S.J. Dixon, K.M. Lemberg, M.R. Lamprecht, R.Skouta, E.M. Zaitsev, C.E. Gleason, et al., Ferroptosis: an iron-dependent form of nonapoptotic cell death, Cell 149(2012) 1060-1072.

RI PT

[9] W.S. Yang, R. SriRamaratnam, M.E. Welsch, K. Shimada, R. Skouta, V.S. Viswanathan, et al., Regulation of ferroptotic cancer cell death by GPX4, Cell 156(2014) 317-331. [10] G.O. Latunde-Dada, Ferroptosis: Role of lipid peroxidation, iron and ferritinophagy, Biochim Biophys Acta 1861(2017) 1893-1900.

SC

[11] E. Lachaier, C. Louandre, C. Godin, Z. Saidak, M. Baert, M. Diouf, et al., Sorafenib induces ferroptosis in human cancer cell lines originating from different solid tumors, Anticancer Res 34(2014) 6417-6422.

M AN U

[12] N. Eling, L. Reuter, J. Hazin, A. Hamacher-Brady, N.R.Brady, Identification of artesunate as a specific activator of ferroptosis in pancreatic cancer cells, Oncoscience 2(2015):517-532. [13] Y. Yamaguchi, T. Kasukabe, S.Kumakura, Piperlongumine rapidly induces the death of human pancreatic cancer cells mainly through the induction of ferroptosis, Int J Oncol 52(2018) 1011-1022.

TE D

[14] Y. Yu, Y. Xie, L. Cao, L. Yang, M. Yang, M.T. Lotze, et al., The ferroptosis inducer erastin enhances sensitivity of acute myeloid leukemia cells to chemotherapeutic agents, Mol Cell Oncol 2(2015) e1054549.

EP

[15] T. Sehm, M. Rauh, K. Wiendieck, M. Buchfelder, I.Y. Eyüpoglu, N.E. Savaskan, Temozolomide toxicity operates in a xCT/SLC7a11 dependent manner and is fostered by ferroptosis, Oncotarget 7(2016) 74630-74647.

AC C

[16] J.L. Roh, E.H. Kim, H.J. Jang, J.Y. Park, D.Shin, Induction of ferroptotic cell death for overcoming cisplatin resistance of head and neck cancer, Cancer Lett 381(2016):96-103. [17] M.L. Liu, D. Sun, T. Li, H. Chen, A systematic review of the immune-regulating and anticancer activities of pseudolaric acid B, Front Pharmacol 8(2017):394. [18] J. Yu, Z. Wang, P. Ren, T. Zhong, Y. Wang, F. Song, et al., Pseudolaric acid B inhibits the secretion of hepatitis B virus, Oncol Rep 37(2017):519-525. [19] J. Tong, S. Yin, Y. Dong, X. Guo, L. Fan, M. Ye, et al., Pseudolaric acid B induces caspase-dependent apoptosis and autophagic cell death in prostate cancer cells, Phytother Res 27(2013) 885-891. [20] X. Gong, M. Wang, Z. Wu, S. Tashiro, S. Onodera, T. Ikejima, Pseudolaric acid B induces

ACCEPTED MANUSCRIPT apoptosis via activation of c-Jun N-terminal kinase and caspase-3 in HeLa cells, Exp Mol Med 36(2004):551-556. [21] J.H. Yu, Q. Cui, Y.Y. Jiang, W. Yang, S. Tashiro, S. Onodera, et al., Pseudolaric acid B induces apoptosis, senescence, and mitotic arrest in human breast cancer MCF-7, Acta Pharmacol Sin 28(2007) 1975-1983.

RI PT

[22] D. Wang, Y. Xin, Y. Tian, W. Li, D. Sun, Y. Yang. Pseudolaric acid B inhibits gastric cancer cell metastasis in vitro and in haematogenous dissemination model through PI3K/AKT, ERK1/2 and mitochondria-mediated apoptosis pathways, Exp Cell Res 352(2017) 34-44.

SC

[23] Q. Sun, Y. Li, The inhibitory effect of pseudolaric acid B on gastric cancer and multidrug resistance via Cox-2/PKC-α/P-gp pathway, PLoS One 9(2014) e107830.

M AN U

[24] V.K. Wong, P. Chiu, S.S. Chung, L.M. Chow, Y.Z.Zhao, B.B. Yang, et al., Pseudolaric acid B, a novel microtubule-destabilizing agent that circumvents multidrug resistance phenotype and exhibits antitumor activity in vivo, Clin Cancer Res 11(2005) 6002-6011. [25] Q. Song, S. Jiang, X. Zhang, C. Pan, C. Lu, J. Peng, et al., Radiosensitivity of human ovarian cancer cells is enhanced by pseudolaric acid B due to the inhibition of the Ras/Raf/ERK signaling pathway, Exp Ther Med 15(2018) 685-690.

TE D

[26] W. Hou, Y. Xie, X. Song, X. Sun, M.T. Lotze, H.J. Zeh 3rd, et al., Autophagy promotes ferroptosis by degradation of ferritin, Autophagy 12(2016) 1425-1428.

EP

[27] S.H. Hong, D.H. Lee, Y.S. Lee, M.J. Jo, Y.A. Jeong, W.T. Kwon, et al., Molecular crosstalk between ferroptosis and apoptosis: emerging role of ER stress-induced p53-independent PUMA expression, Oncotarget 8(2017) 115164-115178.

AC C

[28] Z. Zhou, B. Lu, C. Wang, Z. Wang, T. Luo, M. Piao, et al., RIP1 and RIP3 contribute to shikonin-induced DNA double-strand breaks in glioma cells via increase of intracellular reactive oxygen species, Cancer Lett 390(2017) 77-90. [29] H. Yuan, X. Li, X. Zhang, R. Kang, D. Tang, CISD1 inhibits ferroptosis by protection against mitochondrial lipid peroxidation, Biochem Biophys Res Commun 478(2016) 838-844. [30] R. Skouta, S.J. Dixon, J. Wang, D.E. Dunn, M. Orman, K. Shimada, et al., Ferrostatins inhibit oxidative lipid damage and cell death in diverse disease models, J Am Chem Soc. 2014 Mar 26;136(12) 4551-4556. [31] A. Ayala, M.F. Muñoz, S.Argüelles, Lipid peroxidation: production, metabolism, and signaling mechanisms of malondialdehyde and 4-hydroxy-2-nonenal, Oxid Med Cell Longev 2014(2014) 360438.

ACCEPTED MANUSCRIPT [32] J.M.Gutteridge, Lipid peroxidation initiated by superoxide-dependent hydroxyl radicals using complexed iron and hydrogen peroxide, FEBS Lett 172(1984) 245-249. [33] T. Shono, N. Yokoyama, T. Uesaka, J. Kuroda, R. Takeya, T. Yamasaki, et al., Enhanced expression of NADPH oxidase Nox4 in human gliomas and its roles in cell proliferation and survival, Int J Cancer 123(2008) 787-792.

RI PT

[34] H. Imai, M.Matsuoka, T. Kumagai, T. Sakamoto, T. Koumura, Lipid peroxidation-dependent cell death regulated by GPx4 and Ferroptosis, Curr Top Microbiol Immunol 403(2017) 143-170. [35] Y. Xie, W. Hou, X. Song, Y. Yu, J. Huang, X. Sun, et al., Ferroptosis: process and function, Cell Death Differ 23(2016) 369-379.

SC

[36] D.L. Schonberg, T.E. Miller, Q. Wu, W.A. Flavahan, N.K. Das, J.S. Hale, et al., Preferential iron trafficking characterizes glioblastoma stem-like cells. Cancer Cell 28(2015) 441-455.

M AN U

[37] E. Ooko, M.E. Saeed, O. Kadioglu, S. Sarvi, M. Colak, K. Elmasaoudi, et al., Artemisinin derivatives induce iron-dependent cell death (ferroptosis) in tumor cells, Phytomedicine 22(2015)1045-1054. [38] M. Gao, P. Monian, N. Quadri, R. Ramasamy, X.Jiang, Glutaminolysis and Transferrin Regulate Ferroptosis, Mol Cell 59(2015) 298-308.

TE D

[39] S. Ma, E.S. Henson, Y. Chen, S.B.Gibson, Ferroptosis is induced following siramesine and lapatinib treatment of breast cancer cells, Cell Death Dis 7(2016) e2307.

EP

[40] X. Liu, A.B. Madhankumar, B. Slagle-Webb, J.M. Sheehan, N. Surguladze, J.R.Connor, Heavy chain ferritin siRNA delivered by cationic liposomes increases sensitivity of cancer cells to chemotherapeutic agents, Cancer Res 71(2011) 2240-2249.

AC C

[41] X. Tong, H. Kawabata, H.P.Koeffler, Iron deficiency can upregulate expression of transferrin receptor at both the mRNA and protein level, Br J Haematol 116(2002) 458-464. [42] D.H. Manz, N.L. Blanchette, B.T. Paul, F.M. Torti, S.V.Torti, Iron and cancer: recent insights, Ann N Y Acad Sci 1368(2016)149-161. [43] A.W.Girotti, Lipid hydroperoxide generation, turnover, and effector action in biological systems, J Lipid Res 39(1998)1529-1542. [44] S. Guo, X. Chen, The human Nox4: gene, structure, physiological function and pathological significance, J Drug Target 23(2015) 888-8896. [45] M. Qi, S. Fan, G. Yao, Z. Li, H. Zhou, S. Tashiro, et al., Pseudolaric acid B-induced autophagy contributes to senescence via enhancement of ROS generation and mitochondrial

ACCEPTED MANUSCRIPT dysfunction in murine fibrosarcoma L929 cells, J Pharmacol Sci 121(2013) 200-211. [46] E. Kurtoglu, A. Ugur, A.K. Baltaci, R. Mogolkoc, L. Undar, Activity of neutrophil NADPH oxidase in iron-deficient anemia, Biol Trace Elem Res 96(2003)109-115.

RI PT

[47] H. Zhao, B. Ji, J. Chen, Q. Huang, X. Lu, Gpx 4 is involved in the proliferation, migration and apoptosis of glioma cells, Pathol Res Pract 213(2017) 626-633. [48] R.R. Zhang, K.B. Pointer, J.S. Kuo, Excitotoxic SLC7A11 expression is a marker of poor glioblastoma survival and a potential therapeutic target, Neurosurgery 77(2015) N16-17.

SC

[49] L. Sleire, B.S. Skeie, I.A. Netland, H.E. Førde, E. Dodoo, F. Selheim, et al., Drug repurposing: sulfasalazine sensitizes gliomas to gamma knife radiosurgery by blocking cystine uptake through system Xc-, leading to glutathione depletion, Oncogene 34(2015) 5951-5959.

M AN U

[50] D. Chen, Z. Fan, M. Rauh, M. Buchfelder, I.Y. Eyupoglu, N.Savaskan, ATF4 promotes angiogenesis and neuronal cell death and confers ferroptosis in a xCT-dependent manner. Oncogene, 36(2017) 5593-5608. [51] L. Jiang, N. Kon, T. Li, S.J. Wang, T. Su, H. Hibshoosh, R. Baer, et al., Ferroptosis as a p53-mediated activity during tumour suppression, Nature 520(2015) 57-62.

TE D

[52] G.D. Yao, J. Yang, Q. Li, Y. Zhang, M. Qi, S.M. Fan, et al., Activation of p53 contributes to pseudolaric acid B-induced senescence in human lung cancer cells in vitro, Acta Pharmacol Sin 37(2016) 919-929.

Figure Legends

EP

[53] A.G. Meng, L.L. Jiang, Pseudolaric acid B-induced apoptosis through p53-dependent pathway in human gastric carcinoma cells, J Asian Nat Prod Res 11(2009) 142-152.

AC C

Figure 1. PAB inhibited viability and induced death in glioma cells. (A) Chemical structure of PAB. (B, C, D and E) MTT assay showed that PAB inhibited the viabilities of C6, SHG-44, U251 and U87 glioma cells in a dosage- and time-dependent manner. (F) LDH release assay demonstrated that PAB triggered dosage-dependently glioma cell death. (G) MTT assay proved that treatment with PAB for 48h had little toxic effect on normal cells (HL-7720). The values are expressed as mean±SEM (n=5 per group). *: p < 0.01 versus control group. Figure 2. Iron regulated PAB-induced glioma cell death. (A) PAB increased intracellular ferrous iron in C6, SHG-44, U251 and U87 glioma cells in a dosage- and time-dependent manner. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB at the same concentration for 12h. (B) Pretreatment with deferoxamine (DFO) at 500µmol/L for 1h mitigated PAB-induced increase of intracellular iron. LDH release assay showed that the glioma cell death induced by PAB was inhibited in the presence of DFO (C),

ACCEPTED MANUSCRIPT but was exacerbated by supplement of FAC (D). (E) Western blotting analysis showed that PAB upregulated the expression of TFR, TF, FTH, FTL, and FPN in a time-dependent manner. The values are expressed as mean±SEM (n=5 per group).

SC

RI PT

Figure 3. PAB induced morphological changes in glioma cells. (A) Representative light microscopic images of SHG-44 and U251 glioma cells. Majority of the cells treated with PAB alone became round in shape and smaller in size, which were reversed by pretreatment with DFO, Fer-1 or GKT137831. (B) Representative transmission electronic microscopic images showed that the U251 cell treated with 2µmol/L PAB for 12h had intact membrane with decreased microvilli, shrunken mitochondria with increased membrane density (arrowhead), and normal-sized nucleus free of chromatin condensation or margination (asterisk). (C) Representative fluorescence microscopic images of U251 cells stained with JC-1. (D) Statistical analysis of the red and green fluorescence detected by JC-1. (E) Flow cytometry analysis combined with JC-1 staining showed that PAB decreased the mitochondrial membrane potential.

AC C

EP

TE D

M AN U

Figure 4. Iron mediated PAB-induced accumulation of H2O2 and superoxide. (A) PAB excessively produced MDA in glioma cells in a dosage- and time-dependent manner. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB at the same concentration for 12h. (B) The overproduction of MDA caused by PAB was prevented by pretreatment with Fer-1, Vitamin E, GKT 137831 or GSH, but was promoted by FAC. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB alone; &: p < 0.05 versus the group treated with PAB alone. (C) LDH release assay showed that PAB-induced glioma cell death was mitigated in the presence of Fer-1, Vitamin E or GSH. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB alone. (D) PAB induced accumulation of H2O2 in glioma cells in a dosage- and time-dependent manner. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB at the same concentration for 12h. (E) Observation by fluorescence microscopy showed that the red fluorescence detected by superoxide probe DHE was obviously stronger in the cells treated with 2.0µmol/L PAB for 24h. (F) Statistical analysis of the fluorescence intensity proved that PAB induced over generation of superoxide in glioma cells in a dosage- and time-dependent manner. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB at the same concentration for 12h. (G) PAB-induced increase of superoxide was mitigated by DFO or GKT137831, but was reinforced by FAC. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB alone; (H) PAB-induced accumulation of H2O2 was alleviated in the presence of DFO, GKT137831 or GSH, but was improved by FAC. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB alone; The values are expressed as mean±SEM (n=5 per group). Figure 5. Iron contributed to PAB-induced activation of NADPH oxidase 4. (A) PAB increased the activity of Nox4 in a dosage- and time-dependent manner. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB at the same concentration for 12h. (B)Western blotting revealed that PAB upregulated the expression of Nox4 in a dosage- and time-dependent manner. (C) The activation of Nox4 caused by PAB was prevented by DFO or GKT137831, but was improved by FAC. *: p < 0.01 versus control group; #: p < 0.01 versus the

ACCEPTED MANUSCRIPT

RI PT

group treated with PAB alone. (D) Western blotting analysis proved that the upregulation of Nox4 induced by PAB was inhibited in the presence of FAC, but was enhanced by FAC. (E) LDH release assay showed that PAB-induced glioma cell death was inhibited in the presence of GKT137831. (F) Western blotting demonstrated that GKT137831 prevented PAB-induced expressional upregulation of Nox4. (G) Western blotting showed that knockdown of Nox4 with SiRNA mitigated PAB-induced upregulation of Nox4. (H) LDH release assay revealed that knockdown of Nox4 protected glioma cells against PAB-induced death. (I) PAB-induced accumulation of intracellular H2O2 was prevented when Nox4 was knocked down. (J) Knockdown of Nox4 with SiRNA prevented PAB-induced overproduction of MDA. The values are expressed as mean±SEM (n=5 per group).

TE D

M AN U

SC

Figure 6. GSH depletion contributed to PAB-induced accumulation of H2O2 and lipid peroxides. (A) Western blotting proved that PAB upregulated the expression of GPX4, p53 and phosphor-p53, but downregulated xCT in a time-dependent manner. (B) Western blotting showed that knockdown of GPX4 with SiRNA prevented PAB-induced upregulation of GPX4, but improved the upregulation of TFR, p53 and phospho-p53 and exacerbated the downregulation of xCT. (C) LDH release assay demonstrated that knockdown of GPX4 promoted PAB-induced death in glioma cells. (D) PAB depleted intracellular GSH in a dosage- and time-dependent manner. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB at the same concentration for 12h. (E) PAB triggered cysteine depletion in a dosage- and time-dependent manner. *: p < 0.01 versus control group; #: p < 0.01 versus the group treated with PAB at the same concentration for 12h. The values are expressed as mean±SEM (n=5 per group).

AC C

EP

Figure 7. PAB improved ferrous iron and lipid oxidation in vivo. (A) Representative images of the nude mice with subcutaneous xenograft glioma. The xenograft gliomas in the mouse treated with PAB was significantly smaller than those in the control mouse. (B) PAB inhibited the growth of the xenograft glioma in nude mice. (C) PAB improved dosage-dependently the ferrous iron in the xenograft glioma. (D) Western blotting showed that PAB upregulated the protein levels of TFR, TF, FTH, FTL, Nox4, p53, phospho-p53 and GPX4, but downregulated xCT in the xenograft glioma. (E and F) PAB improved the levels of MDA and H2O2 in a dosage dependent manner (G and H) GSH and cysteine were both depleted by PAB in a concentration-dependent manner. The values are expressed as mean±SEM (n=10 per group). *: p < 0.01 versus control group. Figure 8. Schematic diagram for PAB-induced ferroptosis in glioma cells.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT Highlights for review 1 PAB inhibited the viabilities of glioma cells in vitro and in vivo. 2 The glioma cell death triggered by PAB was decided by intracellular iron and lipid peroxidation. 3 The increase of intracellular iron caused by PAB was related to upregulated transferrin receptor. 4 PAB induced lipid peroxidation via activation of Nox4 in an iron-dependent manner.

AC C

EP

TE D

M AN U

SC

RI PT

5 The depletion of GSH caused by PAB was via p53-regualted inhibition of xCT.