Nrf2 inhibition reverses resistance to GPX4 inhibitor-induced ferroptosis in head and neck cancer

Daiha Shin, Eun Hye Kim, Jaewang Lee, Jong- Lyel Roh

www.elsevier.com

PII: S0891-5849(18)32219-6
DOI: https://doi.org/10.1016/j.freeradbiomed.2018.10.426
Reference: FRB13990

To appear in: Free Radical Biology and Medicine Received date: 8 April 2018
Revised date: 15 September 2018 Accepted date: 13 October 2018

Cite this article as: Daiha Shin, Eun Hye Kim, Jaewang Lee and Jong-Lyel Roh, Nrf2 inhibition reverses resistance to GPX4 inhibitor-induced ferroptosis in head and neck cancer, Free Radical Biology and Medicine, https://doi.org/10.1016/j.freeradbiomed.2018.10.426

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 galley proof before it is published in its final citable 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.

 

Nrf2 inhibition reverses resistance to GPX4 inhibitor-induced ferroptosis in head and neck

cancer

Daiha Shin, Eun Hye Kim, Jaewang Lee, Jong-Lyel Roh*

Department of Otolaryngology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea
* Corresponding author: Department of Otolaryngology, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Republic of Korea. Phone: +82- 2-3010-3965; Fax: +82-2-489-2773. [email protected].
ABSTRACT

Glutathione peroxidase 4 (GPX4) is a regulator of ferroptosis (iron-dependent, non-apoptotic cell death); its inhibition can render therapy-resistant cancer cells susceptible to ferroptosis. However, some cancer cells develop mechanisms protective against ferroptosis; understanding these mechanisms could help overcome chemoresistance. In this study, we investigated the molecular mechanisms underlying resistance to ferroptosis induced by GPX4 inhibition in head and neck cancer (HNC). The effects of two GPX4 inhibitors, (1S, 3R)-RSL3 and ML-162, and of trigonelline were
tested in HNC cell lines, including cisplatin-resistant (HN3R) and acquired RSL3-resistant (HN3-rslR) cells. The effects of the inhibitors and trigonelline, as well as of inhibition of the p62, Keap1, or Nrf2 genes, were assessed by cell viability, cell death, lipid ROS production, and protein expression, and in mouse tumor xenograft models. Treatment with RSL3 or ML-162 induced the ferroptosis of HNC
cells to varying degrees. RSL3 or ML-162 treatment increased the expression of p62 and Nrf2 in chemoresistant HN3R and HN3-rslR cells, inactivated Keap1, and increased expression of the phospho-PERK–ATF4–SESN2 pathway. Transcriptional activation of Nrf2 was associated with resistance to ferroptosis. Overexpression of Nrf2 by inhibiting Keap1 or Nrf2 gene transfection rendered chemosensitive HN3 cells resistant to RSL3. However, Nrf2 inhibition or p62 silencing sensitized HN3R cells to RSL3. Trigonelline sensitized chemoresistant HNC cells to RSL3 treatment in a mouse model transplanted with HN3R. Thus, activation of the Nrf2–ARE pathway contributed to

 

the resistance of HNC cells to GPX4 inhibition, and inhibition of this pathway reversed the resistance to ferroptosis in HNC.
Graphic abstract

A proposed model of Nrf2-induced resistance to GPX4 inhibition-induced ferroptosis in cancer cells. The GPX4 inhibitors RSL3 and ML-162 induce endoplasmic reticulum stress and subsequently p62 expression via the PERK–ATF4–SESN2 pathways. Nrf2 is activated by p62–Keap1 interaction and antioxidant response elements (ARE) related to iron and antioxidant systems is increased, resulting in a decreased labile iron pool, thus contributing to the resistance to ferroptosis.
fx1 Abbreviations:
ARE, antioxidant response element; ATF4, activating transcription factor-4; CI, combination index; FTH1, ferritin heavy chain 1; FPN, ferroportin; GPX4, glutathione peroxidase 4; GSTP1, glutathione S-transferase P; HNC, head and neck cancer; HO-1, hemeoxygenase-1; Keap1, Kelch-like ECH- associated protein 1; LC3, microtubule-associated protein 1A/1B-light chain 3; LIP, labile iron pool; NCOA4, nuclear receptor coactivator 4; NQO1, NAD(P)H dehydrogenase [quinone] 1; Nrf2, Nuclear factor (erythroid-derived 2)-like 2; PERK, protein kinase R-like endoplasmic reticulum kinase; ROS, reactive oxygen species; SESN2, sestrin-2; TrxR1, thioredoxin reductase 1.
Keywords:

head and neck cancer, ferroptosis, GPX4, Nrf2, p62

1.Introduction

Ferroptosis is a recently recognized form of regulated cell death that involves iron accumulation and lipid peroxidation; it is distinct from apoptosis, necroptosis, and autophagic cell death [1]. Glutathione peroxidase (GPX4) is an essential regulator of ferroptosis, acting through the suppression of lipid peroxidation generation [2]. Another key molecule related to ferroptosis is the antiporter system xc– (xCT); this exchanges extracellular cystine for intracellular glutamate as a source of

 

glutathione (GSH), a major cellular antioxidant [3]. Inhibition of xCT and GPX4 can induce the death of cancer cells resistant to conventional chemotherapy or radiotherapy [4]. xCT inhibition induces the depletion of GSH by blocking the uptake of cystine, and it sensitizes cancer cells to chemotherapeutic agents [5, 6]; and GPX4 inhibition renders mesenchymal therapy-tolerant persister cancer cells vulnerable to ferroptotic cancer cell death [7]. Inhibition of the lipid peroxidase pathway that protects against ferroptosis could render therapy-resistant cancer cells susceptible to ferroptotic cell death [8].
Therapy-resistant cancer cells persistently evade cell death, including from apoptosis and ferroptosis. Acyl-CoA synthetase long-chain family member 4 (ACSL4) critically determines cells’ sensitivity to ferroptosis by enriching cellular membranes with long polyunsaturated ω6 fatty acids [9, 10]. Heat shock protein beta-1 (HSPB1) is a negative regulator of ferroptosis; inhibition of HSPB1 phosphorylation increases the anticancer activity of erastin, an inhibitor of system xc– [11]. Another negative regulator of ferroptotic cancer cell death is CDGSH iron sulfur domain 1 (CISD1, also referred to as mitoNEET), which protects against mitochondrial lipid peroxidation [12]. Similarly,
heat shock 70-kDa protein 5 (HSPA5) negatively regulates ferroptosis by protecting cancer cells against lipid peroxidation [13]. Activation of nuclear factor (erythroid-derived 2)-like 2 (Nrf2) protects hepatocellular carcinoma cells against ferroptosis; inhibition of Nrf2 genetically or pharmacologically increases the anticancer activity of sorafenib and erastin [14].
Transcription factor Nrf2 plays a key role in regulating cellular redox homeostasis through binding of its promoter to target genes that contain antioxidant response elements (AREs) [15]. Because the proteasomal activity of Kelch-like ECH-associated protein 1 (Keap1) constantly degrades Nrf2, the inhibition of Keap1 activates the Nrf2–ARE pathway upon oxidative or electrophilic stress [16]. Cancer cells buffer the levels of cellular reactive oxygen species (ROS) by actively upregulating the antioxidant pathways that contribute to cancer therapy resistance [17, 18]. Nrf2 has also been reported to have a protective role against ferroptosis [14]. However, further studies are required to identify the mechanisms underlying resistance to GPX4 inhibition of cancer cells, to facilitate the implementation of new approaches to overcome this resistance. This study identified a mechanism for resistance to GPX4 inhibition in HNC cells, which involved p62 expression and activation of that Nrf2–ARE

 

system. Nrf2 inhibition or p62 silencing sensitized the resistant HNC cells to GPX4 inhibitors in vitro and in vivo.
2.Results

2.1.RSL3 or ML-162 induced the ferroptosis of HNC cells to varying degrees

RSL3, a GPX4 inhibitor, reduced the viability of cisplatin-sensitive HN3 cells in a dose-dependent manner, with the sensitivity greatly decreased in acquired cisplatin-resistant (HN3R) or RSL3- resistant (HN3-rslR) cells (P < 0.01) (Figures 1A and 1C). The same findings were also observed by erastin, an inhibitor of the xCT, when treated in these cells (Figure 1B). Sensitivity to RSL3 treatment differed considerably between the various HNC cell lines and normal cells (Supplementary Fig. S1). RSL3 sensitivity in the RSL3-rslR cells significantly increased by the knockdown of GPX4 gene (Supplementary Fig. S2). In the same line, another GPX4 inhibitor, ML-162, induced HNC cell death to varying degrees, with parental HN3 cells more sensitive and HN3R and HN3-rslR cells less sensitive (P < 0.01) (Supplementary Fig. S3). Propidium iodide staining showed high positivity in HN3 cells, even with low-dose (2 μM) RSL3 treatment, but low positivity in HN3R and HN3-rslR cells (P < 0.01) (Figure 1D). Changes in cellular lipid ROS measured by BODIPY C11 after RSL3 treatment showed a similar trend in these cell lines (Figure 1E). Cell death induced in HNC cells by RSL3 treatment was reversed by co-treatment with an iron chelator deferoxamine, a ferroptosis inhibitor ferrostatin-1, or α-tocopherol, which conferred a pattern typical of ferroptosis (Figure 1F).
2.2.Resistance to ferroptosis was associated with p62 and Nrf2 expression in RSL3-treated HNC cells
RSL3 reduced the expression of GPX4 protein in a dose-dependent manner in both HN3 and HN3R cells (Figure 2A). However, it markedly increased the expression of p62 protein in chemoresistant HN3R cells, but not in chemosensitive HN3 cells, along with the increased expression of Nrf2, LC3-II, and Atg5 proteins, and the decreased expression of Keap1 protein. The accumulation of p62 protein induced by RSL3 was co-localized with LC3 protein in the cytoplasm (Figure 2B and

 

Supplementary Fig. S4). Changes in p62 and Nrf2 protein expression in resistant HN3R cells were also observed following treatment with another GPX4 inhibitor, ML-162 (Supplementary Fig. S5). Inhibition of the p62 gene significantly reduced cell viability and increased cellular lipid ROS levels in HN3R cells; this was reversed by treatment with ferrostatin-1 or α-tocopherol (all P < 0.01) (Figures 2C–2E and Supplementary Fig. S6). In addition, the p62 genetic silencing decreased the expression of Keap1 and the accumulation of LC3-II, but did not change the expression of ACSL4 (Supplementary Fig. S7).
As well as the cytoplasmic accumulation of p62, treatment with RSL3 or ML-162 reduced Keap1, resulting in a marked increase of Nrf2 protein (Figure 3A). The inactivation of Keap1 induced a nuclear accumulation of Nrf2 because of the direct interaction between Keap1 and p62 previously noted [19, 20] (Figures 3B and 3C). The increased transcriptional activity of Nrf2 induced the activation of AREs, including FTH1, FPN, HO-1, NQO1, TrxR1, and GSTP1 (Figures 3D–3G and Supplementary Fig. S8). Treatment of chemosensitive HN3 cells with RSL3 resulted in NCOA4, which acts as a cargo receptor in ferritinophagy [21], being localized in the nucleus and then conferred to autophagosomes in the cytoplasm (Supplementary Fig. S9). In chemoresistant HN3R cells, the NCOA4 did not degrade in the autophagosomes but accumulated, along with FTH1 and other ARE proteins related to iron (Figure 3F). As a result, LIP appeared to increase in HN3 cells but not in HN3R cells (Figure 3G). Further, RSL3 sensitivity in the HN3R cells was significantly increased by the knockdown of NCOA4 gene (Supplementary Fig. S10). FTH1 expression significantly increased along with the NCOA4 gene silencing, whereas LIP was not significantly changed.
We next examined the effect of Nrf2 overexpression on cell growth and viability of chemosensitive HN3 cells. Overexpression of the Nrf2–ARE pathways could be induced by the inhibition of Keap1 or Nrf2 gene transfection. When the Keap1 gene was inhibited by siRNA transfection, Nrf2 and p62 expression increased in HN3 cells with or without RSL3 treatment (Figures 4A and 4B). Overexpression of Nrf2 by Keap1 inhibition or Nrf2 gene transfection recovered the growth and viability of RSL3-sensitive HN3 cells that had been reduced by treatment with RSL3 (Figures 4C and

 

4D). In HN3 cells with Nrf2 gene transfection, RSL3 treatment did not significantly increase cellular lipid ROS levels (P > 0.05) (Supplementary Fig. S11). The cell death and lipid/cytosolic ROS levels were enhanced by a combination of RSL3 with alkaloid trigonelline, an inhibitor of the Nrf2 transcriptional factor [22] (Figure 4F and Supplementary Fig. S11).
We also examined several upstream molecules that affect the activity of p62. Sestrins are upstream molecules that promote the p62-dependent autophagy degradation of Keap1 [23]. Sestrin-2 (SESN2) is also induced by via the activation of the PERK–eIF2α–ATF4 pathway following endoplasmic reticulum stress [24]. Our results supported previous findings that RSL induced the increased expression of phospho-PERK (pPERK), ATF4, and SESN2 proteins, resulting in increases in p62 and Nrf2, in resistant HN3R cells but not in HN3 cells (Supplementary Fig. S12). The preconditioning of endoplasmic reticulum (ER) stress using thapsigarsgin markedly reduced the RSL3 sensitivity in HN3 cells, regardless of GPX4 genetic silencing (Supplementary Fig. S13).
2.3.Inhibition of Nrf2 sensitized chemoresistant HNC cells to RSL3 treatment in vitro and in vivo RSL3 treatment significantly suppressed cell growth in chemoresistant HN3R cells with inhibition
of Nrf2 gene (Figures 5A and 5B), and significantly increased cell death in siNrf2-transfected HN3R cells (P < 0.01) (Figure 5C). In addition, the Nrf2 genetic silencing decreased the expression of p62 and NCOA4 and the accumulation of LC3-II (Supplementary Fig. S14). The same results were found when the cells were treated with trigonelline in combination with RSL3 (Figure 5D). Trigonelline combined with RSL3 significantly increased lipid ROS and cell death in HN3R cells (Figures 5E–5F and Supplementary Fig. S15). Trigonelline also reduced the transcriptional activity of Nrf2, resulting in decreased mRNA levels of AREs, including FTH1 and HMOX1 (Figure 5G).
In mice xenograft models, all the mice survived well during and after cell implantation and treatment with vehicle, RSL3, trigonelline, or RSL3 plus trigonelline. They were euthanized 20 days after treatment. RSL3 or trigonelline alone did not significantly inhibit in vivo tumor growth compared with the vehicle control (P > 0.1) (Figures 6A and 6B). However, tumor growth was significantly suppressed by RSL3 plus trigonelline (P < 0.01). Body weight and daily food intake did

 

not change significantly in the control or treatment groups (P < 0.05) (Figure 6C). The levels of ferrous iron, lipid ROS, and RPA measured in the in vivo tumors were significantly higher in the RSL3 plus trigonelline combination group than the control or other treatment groups (P < 0.01). A histological examination of vital organs did not reveal any significant differences between the groups (Supplementary Fig. S16).
3.Discussion

Cancer chemotherapy is increasingly used in HNC as an organ-preserving treatment strategy [25, 26]. The use of platinum drugs and molecular targeted agents can result in acquired resistance and higher levels of toxicity, contributing to poor treatment outcomes [27]. Recently, it was shown that inhibition of GPX4 renders therapy-resistant cancer cells susceptible to ferroptosis [7, 8]; however, some cancer cells persistently evade this. The present study revealed a mechanism underlying resistance to GPX4 inhibition, which involved p62 expression along with activation of the Nrf2–ARE system. The Keap1–Nrf2 antioxidant system is also involved in the RSL3 resistance mechanism. Nrf2 inhibition or p62 silencing sensitized chemoresistant HNC cells to RSL3. Trigonelline reversed the RSL3-induced resistance to ferroptosis in HNC cells via the inhibition of the Nrf2 system, both in vitro and in vivo. Our results showed that activation of the Nrf2–ARE pathway contributed to the resistance of HNC cells to GPX4 inhibition, and that inhibition of this pathway reversed the resistance to ferroptosis in HNC cells.
This study revealed a mechanism for resistance to GPX4 inhibitors in cancer cells. RSL3-resistant HNC cells showed the expression of p62 with RSL treatment. Cytosolic overexpression of p62 is associated with aggressive phenotypes of various human cancers with poor therapeutic response and survival outcomes [28-30]. p62 plays critical roles at the hub of signaling molecules in multiple cellular function, controlling cell survival and apoptosis as well as autophagy [31]. p62 is assembled in autophagic cargos and phosphorylated in an mTORC1-dependent manner; it subsequently induces the expression of Keap1–Nrf2 antioxidant pathways [32]. Because p62 possesses the KIR binding domain that interacts with Keap1, p62 phosphorylation at Ser349 competitively abrogates the

 

interaction between Nrf2 and Keap1 and degrades Keap1 through selective autophagy [32]. Gene expression of p62/SQSTM1 can also be activated by Nrf2 via a positive feedback loop [20], and the p62–Nrf2 pathway is activated even without Keap1 modification [19]. Our study showed that p62 expression was related to the increased expression of the Nrf2–ARE system via the inactivation of Keap1, and that this was the basis of the mechanism for resistance to GPX4 inhibition. In our study, the inhibition of GPX4 expression by high dose RSL3 or ML-162 treatment or GPX4 genetic silencing significantly increased the sensitivity of GPX4 inhibitors. However, the RSL3-induced cell death might be caused by other mechanisms different from GPX4 inhibition, which would require further examinations.
Activation of the Nrf2–ARE system is affected by upstream pathways and mechanisms that modulate p62. Under oxidative stress conditions, SESN2 is upregulated and activates the antioxidant transcriptional factor Nrf2 [33]. The antioxidant effects of sestrins result from augmenting the p62- dependent autophagy-directed degradation of Keap1 and the consequent activation of Nrf2 [23]. The upregulation of SESN2 promotes Keap1 degradation and Nrf2 activation; this protects the liver from oxidative damage [34]. Activation of the Nrf2–ARE system also upregulates the expression of SESN2, resulting in cytoprotective activity against oxidative stress [35]. Furthermore, SESN2 is upregulated during ER or genotoxic stress and under hypoxia [33]. ER stress induces SESN2, depending on the PERK, via triggering an unfolded protein response and activating ATF4 and Nrf2 transcriptional factors; this protects cells from necroptotic cell death [24, 36, 37]. Our study showed that the p62–
Nrf2 pathway was activated by the expression of SESN2 via induction of the PERK-ATF4 pathway in HNC cells resistant to GPX4 inhibition. In addition, RSL3 appeared to induce ER stress
independently of GPX4 inhibition, as shown in HNC cells with or without GPX4 genetic silencing. Ferroptosis was recently shown to be an autophagic cell death process promoted by the
degradation of ferritin [38, 39]. NCOA4 is highly enriched in the double membrane of autophagosomes, where it forms complexes with autophagy-related protein 8 (ATG8), and serves as a selective cargo receptor for the autophagic turnover of ferritin (ferritinophagy) [21]. NCOA4 knockdown or autophagy blockade inhibits ferritinophagy, abrogates the accumulation of cellular LIP

 

and ROS, and curbs eventual ferroptosis [39]. Conversely, NCOA4 overexpression promotes the degradation of ferritin and thus ferroptotic cell death [38]. During ferroptosis, NCOA4 is degraded, ferritin is released, and the level of bioavailable intracellular labile iron increases; this is reversed with impaired ferritinophagy [38, 39]. Similarly, our study showed that the inhibition of GPX4 in chemoresistant HNC cells resulted in NCOA4 failing to be conferred from the nucleus and degraded in autophagosomes, with the result that there was no increase in LIP levels. However, in our study, the genetic silencing of NCOA4 significantly increased the sensitivity to RSL3 treatment in resistant
HNC cells without the change of LIP, which require further examinations on potential other mechanism.
Nrf2 is the central player in the regulation of antioxidant molecules in cells [40]. It controls cellular antioxidant systems in cancer cells, playing a key role in protecting against intracellular and
environmental stress [41, 42]. Nrf2 is constantly degraded by Keap1, and is activated by the inhibition of Keap1 [16]. In hepatoma cells, activation of the p62–Keap1–Nrf2 antioxidant signaling pathway is a negative regulator of ferroptosis, determining the sensitivity to the ferroptosis inducers that inhibit xCT, such as erastin, sulfasalazine, and sorafenib [14]. Knockdown of p62 promotes the accumulation of Keap1 and enhances Keap1-mediated Nrf2 degradation [14]. In addition, the inhibition of Nrf2 and the AREs HO-1, FTH1, and NQO1 has been shown to significantly enhance the antitumor activity of ferroptosis inducers [14]; however, this was not tested in the context of GPX4 inhibition. The present study revealed that some chemoresistant cancer cells evaded ferroptosis induced by GPX4 inhibition because of activation of the p62–Nrf2–ARE pathway. FTH1 stores iron in the cell, FPN exports iron
to the outside of the cell, and HO-1 influences the iron-related antioxidant function protective against oxidative stress. Our study showed that accumulation of the Nrf2–ARE proteins during ferroptosis reduced the sensitivity of cancer cells to the GPX4 inhibitors RSL3 and ML-162 (Figure 7).
In this study, we investigated ways to overcome the resistance of cancer cells to ferroptosis. The genetic inhibition of Nrf2 or p62 sensitized chemoresistant HNC cells to the inducers of GPX4 inhibition and thus to ferroptosis. The pharmacological inhibition of Nrf2 transcriptional activity using trigonelline reversed RSL3-induced resistance to ferroptosis in HNC cells, both in vitro and in

 

vivo. When co-administered with GPX4 inhibitors, the inhibition of Nrf2 enhanced growth suppression, intracellular ROS accumulation, and ferroptotic cell death. This appears to be consistent with previous findings that the inhibition of Nrf2 reversed the resistance of cisplatin-resistant HNC cells to ferroptosis due to artesunate-induced xCT inhibition [43]. Because the aberrant activation of Nrf2–ARE system is associated with cancer progression and therapeutic resistance, there has been recent focus on the inhibitors in the context of cancer therapy [44]. Inhibition of Nrf2 in combination with inducers of ferroptosis effectively overcomes the resistance of HNC cells to GPX4 inhibition.
In conclusion, the results of study suggest that the p62–Nrf2 pathway is involved in the mechanism underlying resistance to GPX4 inhibition in HNC cells. This pathway is activated by the expression of SESN2 via the induction of the PERK–ATF4 pathway in resistant cancer cells. Inhibition of Nrf2 reverses the resistance of cancer cells to ferroptotic cell death, both in vitro and in vivo. Further preclinical and clinical investigations of Nrf2 inhibition combined with GPX4 inhibition in patients with resistant cancer types should be performed to explore this promising anticancer therapy.
4.Materials and Methods

4.1.Cell lines

The study used an HNC cell line (AMC-HN2–11) established at our hospital [45] and SNU cell lines (SNU-1041, -1066, and -1076) purchased from the Korea Cell Line Bank (Seoul, Republic of Korea). The cell lines were authenticated by short tandem repeat DNA fingerprinting and multiplex polymerase chain reaction (PCR). The cells were cultured in Eagle’s minimum essential medium or Roswell Park Memorial Institute 1640 (Thermo Fisher Scientific, Waltham, MA, USA) with 10%
fetal bovine serum at 37°C in a humidified atmosphere containing 5% CO2. Normal oral keratinocytes (HOK) or fibroblasts (HOF) obtained from patients undergoing oral surgery were used for in vitro cell viability assays. HNC cell lines with acquired chemoresistance against cisplatin (HN3R) and RSL3 (HN3-rslR) were developed from cisplatin-sensitive and RSL3-sensitive parental HN3 cells, with continuous exposure to increasing concentrations of cisplatin and RSL3, respectively. Half-maximal inhibitory concentrations (IC50), determined by cell viability assays, were 3.0 µM in HN3 and 29.7

 

µM in HN3-cisR cells for cisplatin, and 0.48 µM in HN3 and 5.8 µM in HN3-rslR cells for RSL3.

 

4.2.Cell viability and cell death assays

Cell viability and cell death after 72-h exposure to (1S, 3R)-RSL3 (Cayman Chemical, Ann Arbor, MI, USA), ML-162 (Cayman Chemical), trigonelline (Sigma-Aldrich, St. Louis, MO, USA), or combinations of these, erastin (Selleckchem, Houston, TX, USA), or pretreatment of thapsigarsgin (Sigma-Aldrich) were assessed using MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, Sigma-Aldrich), trypan blue exclusion, clonogenic assays, and propidium iodide staining. Control cells were exposed to an equivalent amount of dimethyl sulfoxide (DMSO). MTT assays were performed with the tetrazolium compound for 4 h, followed solubilization buffer for 2 h; absorbance was then measured at 570 nm using a SpectraMax M2 microplate reader (Molecular Devices, Sunnyvale, CA, USA). Trypan blue exclusion involved 0.4% trypan blue staining with counting using a hemocytometer. The clonogenic assays were performed with 0.5% crystal violet solution, counting the number of colonies (>50 cells) after culturing for 14 days. For propidium iodide staining, the sample was washed twice with PBS follow by staining with 2.5 μg/ml propidium iodide (Sigma- Aldrich) in PBS, which was applied to each plate for 30 min. The stained cells were analyzed using a FACSCalibur flow cytometer equipped with CellQuest Pro (BD Biosciences, San Jose, CA, USA) and observed on a ZEISS fluorescent microscope (Oberkochen Germany). The mean fluorescent intensity of each group was normalized to that of the control group.
The interaction of two drugs was considered to be synergistic when they resulted in greater growth suppression than the sum of the growth suppressions induced by each drug alone. The combination index (CI) for drug interaction was scored using software (ComboSyn, Inc., Paramus, NJ, USA) and calculated using the Chou–Talalay method, in which CI < 1 indicated a synergistic interaction, CI = 1 an additive interaction, and CI > 1 an antagonistic interaction [46].
4.3.ROS production measurement

Cellular ROS generation in the supernatant of HNC cell lysates treated for 24 h was measured by

 

adding, for 30 min at 37 °C, either 10 µM 2ʹ,7ʹ-dichlorofluorescein diacetate (Enzo Life Sciences, Farmingdale, NY, USA) for cytosolic ROS or 2 µM C11-BODIPY C11 (Thermo Fisher Scientific, Waltham, MA, USA) for lipid peroxidation. ROS levels were analyzed using a FACSCalibur flow cytometer equipped with CellQuest Pro (BD Biosciences, Franklin Lakes, NJ, USA).
4.4.RNA interference and gene transfection

Cisplatin-sensitive HN3 cells were seeded for the gene silencing of KEAP1, and cisplatin-resistant HN3R cells for the gene silencing of SQSTM1 (p62) and NFE2L2 (Nrf2). After 24 h, the cells were transfected with 10 nmol/L of small interfering RNA (siRNA) that targeted human KEAP1, SQSTM1, GPX4 or NFE2L2, or scrambled control siRNA (TriFECTa® RNAi kits; Integrated DNA Technologies, Coralville, IA, USA). The siRNA-induced gene silencing was confirmed by reverse transcription- quantitative polymerase chain reaction (RT-qPCR) from 1–2 µg total RNA for each sample using SensiFAST™ SYBR® No-ROX Kits (Bioline International, Toronto, Canada) after cDNA synthesis using SensiFAST™ cDNA Synthesis Kits (Bioline International) and western blotting using anti- Keap1, anti-p62, anti-GPX4, and anti-Nrf2 antibodies. To generate cells that stably overexpressed NFE2L2 (Nrf2), HN3 cells were stably transfected with a control plasmid or an Nrf2-expressing plasmid (Transomic, Huntsville, AL, USA). Nrf2 overexpression was confirmed using RT-qPCR and western blotting.
4.5.Immunoblotting

Cells were plated, grown with 70% confluence, and then treated with the indicated drugs. The cells were lysed at 4°C in cell lysis buffer (Cell Signaling Technology, Danvers, MA, USA) with a protease/phosphatase inhibitor cocktail (Cell Signaling Technology). A total of 5–15 µg protein was resolved by SDS-PAGE on 10%–15% gels, transferred to nitrocellulose or polyvinylidene difluoride membranes, and probed with primary and secondary antibodies. The following primary antibodies were used: GPX4 (ab125066, Abcam, Cambridge, MA, USA), p62/SQSTM1 (sc-28359, Santa Cruz Biotechnology, Dallas, TX, USA), Keap1 (10503-2-AP, Proteintech, Chicago, IL, USA), LC3A/B

 

(12741, Cell Signaling Technology), Nrf2 (ab31163, Abcam), ACSL4 (ab38420, Abcam), Atg5 (12994, Cell Signaling Technology), protein kinase R-like endoplasmic reticulum kinase (PERK); 3192, Cell Signaling Technology, activating transcription factor-4 (ATF4); YF-MA10073, AbFrontier, Seoul, Korea, ferritin heavy chain 1 (FTH1); 4393, Cell Signaling Technology, ferroportin (FPN); NBP1-21502SS, Novus Biochemicals, Littleton, CO, USA, hemeoxygenase-1 (HO-1); ab125066, Abcam, NAD(P)H dehydrogenase [quinone] 1 (NQO1); GTX100235, GeneTex, Irvine, CA, USA, glutathione S-transferase P 1 (GSTP1); 3369, Cell Signaling Technology, nuclear receptor coactivator 4 (NCOA4); A302-272A, Bethyl Laboratories, Montgomery, TX, USA, and Sestrin-2 (10795-1-AP, Proteintech). Lamin B1 (ab133741, Abcam) and β-actin (BS6007M, BioWorld, Atlanta, GA, USA) served as nuclear and total loading controls, respectively. All antibodies were diluted to between
1:500 and 1:10000.

 

4.6.Immunofluorescence staining

Cells, treated with the indicated drugs or left untreated, were stained with LC3, p62, LC3 and p62 together, or NCOA4 antibodies. 4′,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific) was used as a nuclear counterstain. The cells were fixed with 3.7% paraformaldehyde in pre-warmed complete medium for 15 min at 37°C. The fixed cells were deparaffinized, rehydrated, and stained with the target antibodies and secondary antibodies. The stained cells were observed using a fluorescent microscope for imaging. The stained cells were also observed on a fluorescent microscope.
4.7.Labile iron pool (LIP) and iron assays

Each sample was seeded in plates and treated with the indicated drugs. After 4 h, the supernatant was removed and the plates were washed twice with Hank’s Balanced Salt Solution (HBSS). The cells were labeled by adding 8 μg/mL Calcein AM (Corning Inc., Corning, NY, USA) in HBSS and were incubated on plates for 30 min at 37 °C. After 30 min, the labeling solution was removed and the cells washed twice with PBS before trypsinization with Trypsin-EDTA. They were then neutralized by adding 4% FBS in HBSS and centrifuged for 3 min at 1500 rpm. The collected cells were washed

 

once with HBSS while vortexing and were centrifuged for 3 min at 1500 rpm. They were then resuspended in 250 μL HBSS while vortexing, and go to FACS. Labile iron levels were analyzed
using a FACSCalibur flow cytometer equipped with CellQuest Pro (BD Biosciences), and ferrous iron levels in the cells or tissue extracts were measured by using an iron assay kit (Sigma-Aldrich).
4.8.Nrf2 transcriptional activity assay

The transcriptional activity of Nrf2 was assayed using a Cignal Antioxidant Response Reporter kit (Qiagen, Valencia, CA, USA), according to the manufacturer’s instructions.
4.9.Tumor xenograft

All animal study procedures were performed in accordance with protocols approved by our institution’s Institutional Animal Care and Use Committee. Ten-week-old athymic BALB/c male nude mice (nu/nu) were purchased from Central Lab Animal Inc. (Seoul, Republic of Korea) and HN3R cells were injected subcutaneously into the flank of each. As soon as gross nodules from the tumor implants were detected, the mice were subjected to one of four treatments: vehicle; RSL3 (100 mg/kg intratumorally twice per week) [2]; trigonelline (50 mg/kg daily via oral administration) [47]; or RSL3 plus trigonelline. Each group included 10 mice. Tumor size and body weight were measured twice a week, and the tumor volume was calculated as (length × width2)/2. After scarification, the tumors were isolated and cellular lipid ROS and ferrous iron levels were measured.
4.10.Statistical analysis

The data are presented as mean  standard deviation. The statistical significance of differences between treatment groups was assessed by the Mann–Whitney U-test or analysis of variance (ANOVA) with Bonferroni post-hoc testing, using IBM® SPSS® Statistics version 24.0 for Windows (IBM Corp., Armonk, NY, USA). Statistical significance was defined as a two-sided P value <0.05.

 

Acknowledgments

This study was supported by a grant (no. 2015R1A2A1A15054540) from the Basic Science Research Program through the National Research Foundation of Korea (NRF), Ministry of Science and ICT, Seoul, Republic of Korea (J.-L. Roh).
Conflicts of interests

The authors declare no conflicts of interest.

Supplementary Figure S1. RSL3 induced head and neck cancer (HNC) cell death to varying degrees. Cell viability was measured in different HNC cell lines, normal human oral keratinocytes (HOK), and oral fibroblasts (HOF) were measured after exposure to various concentrations of RSL3 for 72 h. The error bars represent standard deviation from three replicates.

Supplementary Figure S2. RSL3 sensitivity in RSL3 resistant cells according to the knockdown of GPX4. Acquired RSL3-resistant (HN3-rslR) cells were transfected with siRNA control (siCtr) or siGPX4. The siRNA-induced GPX4 gene silencing was confirmed by western blotting. Propidium iodide-positive cells exposed to RSL3 were stained and counted by fluorescent microscopy and flow cytometry. NT indicates the dimethyl sulfoxide (DMSO) control not treated with RSL3. The error bars represent standard deviation from three replicates. * P < 0.01 relative to the Lipofectamine (Lipo) or siRNA control.

Supplementary Figure S3. ML-162, another GPX4 inhibitor, induced head and neck cancer (HNC) cell death to varying degrees. Cell viability was measured in different HNC cell lines, normal human oral keratinocytes (HOK), and oral fibroblasts (HOF) (A), as well as in parental HN3 cells and acquired cisplatin-resistant (HN3R) and RSL3-resistant (HN3R-rslR) cells (B), after exposure to various concentrations of RSL3 for 72 h. The error bars represent standard deviation from three replicates. * P < 0.01 relative to HN3R and HN3-rslR.

Supplementary Figure S4. Immunofluorescence images of LC3 (green) and p62 (red) HN3 and HN3R cells not exposed to RSL3, as the controls of Figure 2B. DAPI was used as a nuclear counterstain.

 

Supplementary Figure S5. Western blot analyses of GPX4, p62, Nrf2, and Keap1 in HN3 and HN3R cells exposed to various concentrations of ML-162 for 24 h. β-actin was used as a loading control.

Supplementary Figure S6. RSL3 sensitivity in HN3R cells with or without the knockdown of p62. Propidium iodide-positive cells exposed to RSL3 were stained and counted by fluorescent microscopy and flow cytometry. NT indicates the dimethyl sulfoxide control not treated with RSL3. The error bars represent standard deviation from three replicates. * P < 0.01 relative to the Lipo or siRNA control.

Supplementary Figure S7. Western blot analyses of GPX4, Keap1, ACSL4, LC3, and p62 in HN3R transfected with siRNA control (siCtr) or sip62 and then exposed to various concentrations of RSL3 for 24 h. β-actin was used as a loading control.

Supplementary Figure S8. Immunofluorescence images of LC3 (green) and NCOA4 (red) in HN3 and HN3R cells exposed to dimethyl sulfoxide control (NT) or 8 μM RSL3. DAPI was used as a nuclear counterstain.

Supplementary Figure S9. Western blot analyses of NQO1, TrxR1, and GSTP1 in HN3 and HN3R cells exposed to various concentrations of RSL3 for 24 h.

Supplementary Figure S10. RSL3 sensitivity and Western blot analyses in HN3R cells transfected with siRNA control (siCtr) or siNCOA4. (A) Propidium iodide-positive cells exposed to RSL3 were stained and counted by fluorescent microscopy and flow cytometry. NT indicates the dimethyl sulfoxide control not treated with RSL3. The error bars represent standard deviation from three replicates. * P < 0.01 relative to the Lipo or siRNA control. (B) Western blot analyses of GPX4, ACSL4, LC3, FTH1, and NCOA4 in HN3R transfected with siRNA control (siCtr) or siNCOA4 and then exposed to various concentrations of RSL3 for 24 h. β-actin was used as a loading control. (C) LIP measured at 4 hours after exposure to 4 μM RSL3.

Supplementary Figure S11. Changes in cellular lipid (A) and cytosolic (B) ROS levels of Nrf2- overexpressed HN3 cells exposed to 4 μM RSL3 and/or 0.3 mM trigonelline (Trig) for 24 h. NT indicates the DMSO control not treated with RSL3 or Trig. The error bars represent standard

 

deviation from three replicates. * P < 0.01 compared between Trig plus RSL3 and the other groups.

Supplementary Figure S12. Western blot analyses of phospho-PERK (pPERK), ATF4, SESN2, p62 and Nrf2 in HN3 and HN3R cells exposed to 8 μM RSL3 for 24 h.

Supplementary Figure S13. RSL3 sensitivity in RSL3-senstivie HN3 cells with or without thapsigarsgin. The cells were pretreated with or without 10 nM thapsigarsgin (Thap), an inducer of endoplasmic reticulum stress, and then exposed to different concentrations of RSL3. (A) Propidium iodide-positive cells were stained and counted by fluorescent microscopy and flow cytometry. NT indicates the dimethyl sulfoxide control not treated with RSL3. The error bars represent standard deviation from three replicates. * P < 0.01 between absence and presence of thapsigarsgin pretreatment. (B) Western blot analyses of pPERK, ATF4, SESN2, and p62 in HN3 cells transfected with siRNA control (siCtr) or siGPX4 and then exposed to 10 nM Thap.

Supplementary Figure S14. Western blot analyses of GPX4, NCOA4, p62, LC3, and Nrf2 in HN3R transfected with siRNA control (siCtr) or siNrf2 and then exposed to 4 μM RSL3 for 24 h. β-actin was used as a loading control.

Supplementary Figure S15. Effects on cell death of Nrf2 pharmacological inhibition induced by trigonelline. The cells were untreated or treated with 4 μM RSL3 with or without 0.3 mM trigonelline (Trig) for 72 h, and then were stained with propidium iodide. The error bars represent standard deviation from three replicates. * P < 0.01 compared between the groups.

Supplementary Figure S16. There was no significant tissue damage in mice treated with RSL3, trigonelline (Trig), or RSL3 + Trig. Major organs were harvested from mice treated with vehicle, RSL3, Trig, or RSL3 + Trig and stained with hematoxylin and eosin.

 

References

[1]S.J. Dixon, K.M. Lemberg, M.R. Lamprecht, R. Skouta, E.M. Zaitsev, C.E. Gleason, D.N. Patel, A.J. Bauer, A.M. Cantley, W.S. Yang, B. Morrison, 3rd, B.R. Stockwell, Ferroptosis: an iron- dependent form of nonapoptotic cell death, Cell 149(5) (2012) 1060-72.

 

[2]W.S. Yang, R. SriRamaratnam, M.E. Welsch, K. Shimada, R. Skouta, V.S. Viswanathan, J.H. Cheah, P.A. Clemons, A.F. Shamji, C.B. Clish, L.M. Brown, A.W. Girotti, V.W. Cornish, S.L. Schreiber, B.R. Stockwell, Regulation of ferroptotic cancer cell death by GPX4, Cell 156(1-2) (2014) 317-331.
[3]S.J. Dixon, D.N. Patel, M. Welsch, R. Skouta, E.D. Lee, M. Hayano, A.G. Thomas, C.E. Gleason, N.P. Tatonetti, B.S. Slusher, B.R. Stockwell, Pharmacological inhibition of cystine-glutamate exchange induces endoplasmic reticulum stress and ferroptosis, Elife 3 (2014) e02523.
[4]Y. Xie, W. Hou, X. Song, Y. Yu, J. Huang, X. Sun, R. Kang, D. Tang, Ferroptosis: process and function, Cell Death Differ 23(3) (2016) 369-79.
[5]M. Yoshikawa, K. Tsuchihashi, T. Ishimoto, T. Yae, T. Motohara, E. Sugihara, N. Onishi, T. Masuko, K. Yoshizawa, S. Kawashiri, M. Mukai, S. Asoda, H. Kawana, T. Nakagawa, H. Saya, O. Nagano, xCT inhibition depletes CD44v-expressing tumor cells that are resistant to EGFR- targeted therapy in head and neck squamous cell carcinoma, Cancer Res 73(6) (2013) 1855-66.
[6]D.S. Liu, C.P. Duong, S. Haupt, K.G. Montgomery, C.M. House, W.J. Azar, H.B. Pearson, O.M. Fisher, M. Read, G.R. Guerra, Y. Haupt, C. Cullinane, K.G. Wiman, L. Abrahmsen, W.A. Phillips, N.J. Clemons, Inhibiting the system xC(-)/glutathione axis selectively targets cancers with mutant-p53 accumulation, Nat Commun 8 (2017) 14844.
[7]M.J. Hangauer, V.S. Viswanathan, M.J. Ryan, D. Bole, J.K. Eaton, A. Matov, J. Galeas, H.D. Dhruv, M.E. Berens, S.L. Schreiber, F. McCormick, M.T. McManus, Drug-tolerant persister cancer cells are vulnerable to GPX4 inhibition, Nature 551(7679) (2017) 247-250.
[8]V.S. Viswanathan, M.J. Ryan, H.D. Dhruv, S. Gill, O.M. Eichhoff, B. Seashore-Ludlow, S.D. Kaffenberger, J.K. Eaton, K. Shimada, A.J. Aguirre, S.R. Viswanathan, S. Chattopadhyay, P. Tamayo, W.S. Yang, M.G. Rees, S. Chen, Z.V. Boskovic, S. Javaid, C. Huang, X. Wu, Y.Y. Tseng, E.M. Roider, D. Gao, J.M. Cleary, B.M. Wolpin, J.P. Mesirov, D.A. Haber, J.A. Engelman, J.S. Boehm, J.D. Kotz, C.S. Hon, Y. Chen, W.C. Hahn, M.P. Levesque, J.G. Doench, M.E. Berens, A.F. Shamji, P.A. Clemons, B.R. Stockwell, S.L. Schreiber, Dependency of a therapy-resistant state of cancer cells on a lipid peroxidase pathway, Nature 547(7664) (2017) 453-457.
[9]H. Yuan, X. Li, X. Zhang, R. Kang, D. Tang, Identification of ACSL4 as a biomarker and contributor of ferroptosis, Biochemical and biophysical research communications 478(3) (2016) 1338-43.
[10]S. Doll, B. Proneth, Y.Y. Tyurina, E. Panzilius, S. Kobayashi, I. Ingold, M. Irmler, J. Beckers, M. Aichler, A. Walch, H. Prokisch, D. Trumbach, G. Mao, F. Qu, H. Bayir, J. Fullekrug, C.H. Scheel, W. Wurst, J.A. Schick, V.E. Kagan, J.P. Angeli, M. Conrad, ACSL4 dictates ferroptosis sensitivity by shaping cellular lipid composition, Nat Chem Biol 13(1) (2017) 91-98.
[11]X. Sun, Z. Ou, M. Xie, R. Kang, Y. Fan, X. Niu, H. Wang, L. Cao, D. Tang, HSPB1 as a novel regulator of ferroptotic cancer cell death, Oncogene 34(45) (2015) 5617-25.
[12]H. Yuan, X. Li, X. Zhang, R. Kang, D. Tang, CISD1 inhibits ferroptosis by protection against mitochondrial lipid peroxidation, Biochemical and biophysical research communications 478(2) (2016) 838-44.
[13]S. Zhu, Q. Zhang, X. Sun, H.J. Zeh, 3rd, M.T. Lotze, R. Kang, D. Tang, HSPA5 Regulates Ferroptotic Cell Death in Cancer Cells, Cancer Res 77(8) (2017) 2064-2077.
[14]X. Sun, Z. Ou, R. Chen, X. Niu, D. Chen, R. Kang, D. Tang, Activation of the p62-Keap1-NRF2 pathway protects against ferroptosis in hepatocellular carcinoma cells, Hepatology 63(1) (2016) 173-84.
[15]J.D. Hayes, A.T. Dinkova-Kostova, The Nrf2 regulatory network provides an interface between redox and intermediary metabolism, Trends Biochem Sci 39(4) (2014) 199-218.

 

[16]A. Kobayashi, M.I. Kang, Y. Watai, K.I. Tong, T. Shibata, K. Uchida, M. Yamamoto, Oxidative and electrophilic stresses activate Nrf2 through inhibition of ubiquitination activity of Keap1, Mol Cell Biol 26(1) (2006) 221-9.
[17]M. Diehn, R.W. Cho, N.A. Lobo, T. Kalisky, M.J. Dorie, A.N. Kulp, D. Qian, J.S. Lam, L.E. Ailles, M. Wong, B. Joshua, M.J. Kaplan, I. Wapnir, F.M. Dirbas, G. Somlo, C. Garberoglio, B. Paz, J. Shen, S.K. Lau, S.R. Quake, J.M. Brown, I.L. Weissman, M.F. Clarke, Association of reactive oxygen species levels and radioresistance in cancer stem cells, Nature 458(7239) (2009) 780-3.
[18]G.M. DeNicola, F.A. Karreth, T.J. Humpton, A. Gopinathan, C. Wei, K. Frese, D. Mangal, K.H. Yu, C.J. Yeo, E.S. Calhoun, F. Scrimieri, J.M. Winter, R.H. Hruban, C. Iacobuzio-Donahue, S.E. Kern, I.A. Blair, D.A. Tuveson, Oncogene-induced Nrf2 transcription promotes ROS detoxification and tumorigenesis, Nature 475(7354) (2011) 106-9.
[19]M. Komatsu, H. Kurokawa, S. Waguri, K. Taguchi, A. Kobayashi, Y. Ichimura, Y.S. Sou, I. Ueno, A. Sakamoto, K.I. Tong, M. Kim, Y. Nishito, S. Iemura, T. Natsume, T. Ueno, E. Kominami, H. Motohashi, K. Tanaka, M. Yamamoto, The selective autophagy substrate p62 activates the stress responsive transcription factor Nrf2 through inactivation of Keap1, Nat Cell Biol 12(3) (2010) 213-23.
[20]A. Jain, T. Lamark, E. Sjottem, K.B. Larsen, J.A. Awuh, A. Overvatn, M. McMahon, J.D. Hayes, T. Johansen, p62/SQSTM1 is a target gene for transcription factor NRF2 and creates a positive feedback loop by inducing antioxidant response element-driven gene transcription, J Biol Chem 285(29) (2010) 22576-91.
[21]J.D. Mancias, X. Wang, S.P. Gygi, J.W. Harper, A.C. Kimmelman, Quantitative proteomics identifies NCOA4 as the cargo receptor mediating ferritinophagy, Nature 509(7498) (2014) 105- 9.
[22]A. Arlt, S. Sebens, S. Krebs, C. Geismann, M. Grossmann, M.L. Kruse, S. Schreiber, H. Schafer, Inhibition of the Nrf2 transcription factor by the alkaloid trigonelline renders pancreatic cancer cells more susceptible to apoptosis through decreased proteasomal gene expression and proteasome activity, Oncogene 32(40) (2013) 4825-35.
[23]S.H. Bae, S.H. Sung, S.Y. Oh, J.M. Lim, S.K. Lee, Y.N. Park, H.E. Lee, D. Kang, S.G. Rhee, Sestrins activate Nrf2 by promoting p62-dependent autophagic degradation of Keap1 and prevent oxidative liver damage, Cell Metab 17(1) (2013) 73-84.
[24]H.J. Kim, Y. Joe, S.K. Kim, S.U. Park, J. Park, Y. Chen, J. Kim, J. Ryu, G.J. Cho, Y.J. Surh, S.W. Ryter, U.H. Kim, H.T. Chung, Carbon monoxide protects against hepatic steatosis in mice by inducing sestrin-2 via the PERK-eIF2alpha-ATF4 pathway, Free Radic Biol Med 110 (2017) 81- 91.
[25]M. Kong, S.E. Hong, Tumor regression patterns based on follow-up duration in patients with head and neck squamous cell carcinoma treated with radiotherapy or chemoradiotherapy, Clin Exp Otorhinolaryngol 8(4) (2015) 416-21.
[26]J.D. Suh, J.H. Cho, Trends in Head and Neck Cancer in South Korea Between 1999 and 2012, Clin Exp Otorhinolaryngol 9(3) (2016) 263-9.
[27]N. Denaro, M.C. Merlano, E.G. Russi, Follow-up in Head and Neck Cancer: Do More Does It Mean Do Better? A Systematic Review and Our Proposal Based on Our Experience, Clin Exp Otorhinolaryngol 9(4) (2016) 287-297.
[28]P. Rolland, Z. Madjd, L. Durrant, I.O. Ellis, R. Layfield, I. Spendlove, The ubiquitin-binding protein p62 is expressed in breast cancers showing features of aggressive disease, Endocr Relat Cancer 14(1) (2007) 73-80.

 

[29]J.M. Park, S. Huang, T.T. Wu, N.R. Foster, F.A. Sinicrope, Prognostic impact of Beclin 1, p62/sequestosome 1 and LC3 protein expression in colon carcinomas from patients receiving 5- fluorouracil as adjuvant chemotherapy, Cancer Biol Ther 14(2) (2013) 100-7.
[30]J.L. Liu, F.F. Chen, J. Lung, C.H. Lo, F.H. Lee, Y.C. Lu, C.H. Hung, Prognostic significance of p62/SQSTM1 subcellular localization and LC3B in oral squamous cell carcinoma, Br J Cancer 111(5) (2014) 944-54.
[31]J. Moscat, M.T. Diaz-Meco, p62 at the crossroads of autophagy, apoptosis, and cancer, Cell 137(6) (2009) 1001-4.
[32]Y. Ichimura, S. Waguri, Y.S. Sou, S. Kageyama, J. Hasegawa, R. Ishimura, T. Saito, Y. Yang, T. Kouno, T. Fukutomi, T. Hoshii, A. Hirao, K. Takagi, T. Mizushima, H. Motohashi, M.S. Lee, T. Yoshimori, K. Tanaka, M. Yamamoto, M. Komatsu, Phosphorylation of p62 activates the Keap1- Nrf2 pathway during selective autophagy, Mol Cell 51(5) (2013) 618-31.
[33]M. Pasha, A.H. Eid, A.A. Eid, Y. Gorin, S. Munusamy, Sestrin2 as a Novel Biomarker and Therapeutic Target for Various Diseases, Oxid Med Cell Longev 2017 (2017) 3296294.
[34]S.J. Kim, K.M. Kim, J.H. Yang, S.S. Cho, J.Y. Kim, S.J. Park, S.K. Lee, S.K. Ku, I.J. Cho, S.H. Ki, Sestrin2 protects against acetaminophen-induced liver injury, Chem Biol Interact 269 (2017) 50-58.
[35]B.Y. Shin, S.H. Jin, I.J. Cho, S.H. Ki, Nrf2-ARE pathway regulates induction of Sestrin-2 expression, Free Radic Biol Med 53(4) (2012) 834-41.
[36]B. Ding, A. Parmigiani, A.S. Divakaruni, K. Archer, A.N. Murphy, A.V. Budanov, Sestrin2 is induced by glucose starvation via the unfolded protein response and protects cells from non- canonical necroptotic cell death, Sci Rep 6 (2016) 22538.
[37]H.W. Park, H. Park, S.H. Ro, I. Jang, I.A. Semple, D.N. Kim, M. Kim, M. Nam, D. Zhang, L. Yin, J.H. Lee, Hepatoprotective role of Sestrin2 against chronic ER stress, Nat Commun 5 (2014) 4233.
[38]W. Hou, Y. Xie, X. Song, X. Sun, M.T. Lotze, H.J. Zeh, 3rd, R. Kang, D. Tang, Autophagy promotes ferroptosis by degradation of ferritin, Autophagy 12(8) (2016) 1425-8.
[39]M. Gao, P. Monian, Q. Pan, W. Zhang, J. Xiang, X. Jiang, Ferroptosis is an autophagic cell death process, Cell Res 26(9) (2016) 1021-32.
[40]M.B. Sporn, K.T. Liby, NRF2 and cancer: the good, the bad and the importance of context, Nat Rev Cancer 12(8) (2012) 564-71.
[41]C. Gorrini, I.S. Harris, T.W. Mak, Modulation of oxidative stress as an anticancer strategy, Nat Rev Drug Discov 12(12) (2013) 931-47.
[42]M.K. Baek, H. Lee, K.O. Kim, H.J. Kwon, M.H. Chung, H.M. Park, J.H. Woo, D.Y. Kim, Age- Related Changes in Nuclear Factor Erythroid 2-Related Factor 2 and Reactive Oxygen Species and Mitochondrial Structure in the Tongues of Fischer 344 Rats, Clin Exp Otorhinolaryngol 10(4) (2017) 357-362.
[43]J.L. Roh, E.H. Kim, H. Jang, D. Shin, Nrf2 inhibition reverses the resistance of cisplatin-resistant head and neck cancer cells to artesunate-induced ferroptosis, Redox Biol 11 (2017) 254-262.
[44]J. Zhu, H. Wang, F. Chen, J. Fu, Y. Xu, Y. Hou, H.H. Kou, C. Zhai, M.B. Nelson, Q. Zhang, M.E. Andersen, J. Pi, An overview of chemical inhibitors of the Nrf2-ARE signaling pathway and their potential applications in cancer therapy, Free Radic Biol Med 99 (2016) 544-556.
[45]S.Y. Kim, K.C. Chu, H.R. Lee, K.S. Lee, T.E. Carey, Establishment and characterization of nine new head and neck cancer cell lines, Acta Otolaryngol 117(5) (1997) 775-84.
[46]T.C. Chou, Drug combination studies and their synergy quantification using the Chou-Talalay method, Cancer Res 70(2) (2010) 440-6.

 

[47]J.L. Roh, H. Jang, E.H. Kim, D. Shin, Targeting of the Glutathione, Thioredoxin, and Nrf2 Antioxidant Systems in Head and Neck Cancer, Antioxid Redox Signal 27(2) (2017) 106-114.
Figure 1. RSL3 induced head and neck cancer cell death to varying degrees. (A–C) Viability and clonogenic assays of parental HN3 cells and acquired cisplatin-resistant (HN3R) and RSL3-resistant (HN3R-rslR) cells that had been exposed to different concentrations of RSL3 or erastin for 72 h. (D) Propidium iodide-positive cells exposed to RSL3 were stained and counted by fluorescent microscopy and flow cytometry. NT indicates the dimethyl sulfoxide control not treated with RSL3. (E) Measurement of cellular lipid ROS levels by using BODIPY C11 after exposure to RSL3 for 24 h. (F) Propidium iodide-positive HN3 cells, and their ATP levels, after exposure to 2 μM RSL3. The cells were co-treated with deferoxamine (DFO, 100 μM), ferrostatin-1 (Fer-1, 1 μM), or alpha-tocopherol (αTP, 1 mM). The error bars represent standard deviation from three replicates. * P < 0.01 relative to the control or the differently treated groups.

Figure 2. Resistance to RSL3 was associated with p62 expression in RSL3-treated head and neck cancer cells. (A, B) Western blot and immunofluorescence images of HN3 and HN3R cells exposed to various concentrations of RSL3 for 24 h. β-actin was used as a loading control and DAPI as a nuclear counterstain. (C) Silencing of the p62 gene (SQSTM1) in HN3R. * P < 0.01 relative to siRNA control (siCtr). (D) Changes in cell number and cellular lipid ROS levels of HN3R cells exposed to 2 μM RSL3. The error bars represent standard deviation from three replicates. * P < 0.01 compared between groups.

Figure 3. RSL3 or ML-162 increased Nrf2 expression levels during ferroptosis in resistant head and neck cancer cells. (A) Changes in Nrf2, Keap1, and p62 protein levels after exposure of HN3R cells to dimethyl sulfoxide control (NT), 8 μM RSL3, or 8 μM ML-162 for 24 h. * P < 0.01 relative to control. (B) Nrf2 mRNA levels assessed in HN3R cells exposed to 8 μM RSL3 or ML-162 for 24 h. (C, D) Nrf2 expression in nuclear extracts and Nrf2 transcriptional activity in HN3R cells treated with 8 μM RSL3 for 24 h. * P < 0.01 between the groups. (F) Western blots of antioxidant response elements (ARE) related to iron and NCOA4 in HN3 and HN3R cells exposed to various concentrations of

 

RSL3 for 24 h. (G) Labile iron pool (LIP) measured in HN3, HN3R, and HN3-rslR cells at different times after exposure to 8 μM RSL3. The error bars represent standard deviation from three replicates. * P < 0.01 compared between the groups.

Figure 4. Nrf2 activation contributed to the resistance to ferroptotic cancer cell death. (A–C)

Cisplatin-sensitive HN3 cells were transfected with siRNA control (siCtr) or siKeap1. The mRNA (A), proteins (B), and change in number of cells (C) were measured either without treatment or after transfection with 4 μM RSL3. (D, F) Effects of Nrf2 overexpression induced by Nrf2 gene
transfection on RSL3-induced changes in cell viability and propidium iodide-positive cell death. The cells were untreated or treated with 2 μM or 4 μM RSL3, with or without 0.3 mM trigonelline (Trig). The error bars represent standard deviation from three replicates. * P < 0.01 compared between the groups.

Figure 5. Nrf2 inhibition sensitized head and neck cancer cells resistant to RSL3 treatment. (A–C) Cisplatin-resistant HN3R cells were transfected with siRNA control (siCtr) or siNrf2. The mRNA (A), change in number of cells (B), and number of propidium iodide-positive cells (C) were measured either in untreated cells or after transfection with 2 μM or 4 μM RSL3. * P < 0.01 compared between the groups. (D–G) Effects of the pharmacological inhibition Nrf2 induced by trigonelline on cell viability, cell death, lipid ROS levels, and mRNA levels of FTH1 and HMOX1. The cells were either untreated or treated with various concentrations of RSL3, with or without 0.15 mM or 0.3 mM trigonelline (Trig). The error bars represent standard deviation from three replicates. * P < 0.01 compared between the groups.

Figure 6. The pharmacological inhibition of Nrf2 sensitized head and neck cancer cells to RSL3 treatment in vivo. (A–C) Tumor growth and weights and body weight changes after the transplantation of HN3R in nude mice. The mice received one of four treatments: vehicle control, RSL3, trigonelline (Trig), or RSL3 + Trig. (D–F) Measurements of ferrous iron, lipid ROS, and rhodamine B-[(1,10- phenanthroline-5-yl)-aminocarbonyl]benzyl ester (RPA) in tumor tissues with the different treatments. The error bars represent the standard deviation. ** P < 0.01 relative to control or other treatment

 

groups.

Figure 7. A proposed model of Nrf2-induced resistance to GPX4 inhibitor-induced ferroptosis in cancer cells. The GPX4 inhibitors RSL3 and ML-162 induce endoplasmic reticulum stress and subsequently p62 expression via the PERK–ATF4–SESN2 pathways. Nrf2 is activated by p62–Keap1 interaction and antioxidant response elements (ARE) related to iron and antioxidant systems is increased, resulting in a decreased labile iron pool, thus contributing to the resistance to ferroptosis.

 

Highlights

 RSL3 and ML-162 inhibited GPX4, inducing the ferroptotic death of head and neck cancer (HNC)

cells to varying degrees.

 Resistance to GPX4 inhibition is related to the increased expression of p62.

 The p62–Keap1–Nrf2 antioxidant system is also involved in the RSL3 resistance mechanism. Nrf2 inhibition sensitized chemoresistant HNC cells to RSL3 treatment.
 Trigonelline reversed RSL3-induced resistance to ferroptosis in HNC via inhibition of the Nrf2

system, both in vitro and in vivo.ML162