HN6 cells were provided by the Affiliated Stomatological Hospital of Shanxi Medical University. CAL27 cells were purchased from the American Type Culture Collection (ATCC), and 293T cells were obtained from Wuhan BioEagle Biological Science & Technology Co., Ltd. All cell lines were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (BL301A, Biosharp) containing 1% penicillin-streptomycin solution (BL505A, Biosharp) and 10% fetal bovine serum (FBS) (FSP500, ExCell), and maintained at 37°C in an incubator with 5% CO₂.

Capsaicin (HY-10448) and Ferrostatin-1 (HY-100579) were purchased from MedChemExpress (Shanghai, China). The antibodies included ACSL4 (22401-1-AP, Proteintech, 1:8000), GPX4 (30388-1-AP, Proteintech, 1:1000), SLC7A11 (A2413, ABclonal, 1:1000), BECN1 (11306-1-AP, Proteintech, 1:1000), p-BECN1 (Ser90, AP1254, ABclonal, 1:1000), AMPK (18167-1-AP, Proteintech, 1:1000), p-AMPK (Thr172, AF3423, Affinity, 1:1000), β-actin as an internal control (AC038, ABclonal, 1:10000) and HRP-conjugated goat anti-rabbit IgG secondary antibody (BS13278, Bioworld, 1:100000).

HN6 and CAL27 cells were seeded in 96-well plates at a density of 1 × 10³ cells per well and cultured until reaching 70% confluence. After treatment with CAP at a concentration of 150 μM for 24 hours, 10 μL Cell Counting Kit-8 (CCK-8) reagent (28) (MA0218, Dalian Meilun Biology Technology Co. Ltd) was added to each well. The optical density (OD) at 450 nm was measured using a microplate reader (MR-96A, Shenzhen Mindray) to assess cell viability.

HN6 and CAL27 cells were initially cultured at a density of 5 × 10⁶ cells in T25 flasks. Once the cells adhered, they were treated with 150 μM CAP for 24 hours. then the cells were digested and collected, fixed with 2.5% glutaraldehyde at 4°C for 6 hours. Following fixation, the cells were dehydrated through a graded series of ethanol and acetone solutions. Subsequently, the dehydrated samples were embedded in resin and sectioned. The sections were stained with a solution of 3% uranyl acetate and 0.4% lead citrate before being examined using a transmission electron microscope (10) (JEM-1400 Plus, JEOL, Japan).

HN6 and CAL27 cells were seeded in 6-well plates at a density of 1×10⁶ per well, followed by treatment with 150 μM CAP, with or without 1 μM Fer-1. MDA levels were measured according to the instructions of the MDA Assay Kit (29) (RXWB0005-96, Ruixin Biotech), and the absorbance was read at 532 nm and 600 nm using a microplate reader (MR-96A, Shenzhen Mindray).

HN6 and CAL27 cells were seeded in 6-well plates at a density of 1×10⁶ per well. Subsequently, they were treated with 150 μM CAP, with or without 1 μM Fer-1. The levels of Fe²⁺ were measured according to the instructions of the Fe²⁺ Assay Kit (26) (RXWB0464-96, Ruixin Biotech), and the absorbance was recorded at 562 nm using a microplate reader (MR-96A, Shenzhen Mindray).

HN6 and CAL27 cells were seeded in 6-well plates at a density of 1×10⁶ per well. Then they were treated with 150 μM CAP, with or without 1 μM Fer-1. GSH levels were detected following the instruction of the GSH Assay Kit (30) (A006-2-1, Nangjing Jiancheng Bioengineering). The absorbance was read at 405 nm using a microplate reader (MR-96A, Shenzhen Mindray).

HN6 and CAL27 cells were inoculated into 6-well plates at a density of 1×10⁶ per well and treated with 150 μM CAP for 24 h, with or without 1 μM Fer-1. After that, the 2’,7’-Dichlorodihydrofluorescein diacetate (31)(DCFH-DA) (S0033S, Beyotime Biotechnology) was added and incubated at 37°C for 30 min, then observed under a fluorescence microscope (MR-96A, Shenzhen Mindray).

Glutamate release levels were detected following the instruction of the Glutamate Content Assay Kit (32) (BC5215, Solarbio). The absorbance was read at 450 nm using a microplate reader (MR-96A, Shenzhen Mindray).

The cultured HN6 and CAL27 cells were harvested and lysed to obtain total protein. The immunoprecipitation procedure was carried out according to the kit instructions (33) (P2179S, Beyotime Biotechnology), using specific antibodies and protein lysates. The precipitated proteins were then collected using protein A/G magnetic beads, followed by thorough washing of the beads. Subsequently, the proteins were eluted by boiling in SDS sample buffer, and Western blot analysis was performed.

The lentiviral vector pLenti-shRNA-GFP-Puro, along with the packaging plasmids PSPAX-2 and PMD2G, were co-transfected into 293T cells. The viral supernatant was collected and filtered through a 0.22 μm filter for subsequent infection. HN6 cells were seeded in a 6-well plate and cultured for 24 hours before viral infection. After 72 hours post-infection, puromycin was added to select infected cells. The knockdown efficiency of AMPK was then confirmed by Western blot analysis (17). The sequence of shRNA was as follow: 5’-GGA AGT TCT CAG CTG TCT TTA-3’.

HN6 and CAL27 cells were cultured in a 6-well plate and treated with CAP at a concentration of 150 μM. Protein extraction was carried out using RIPA lysis buffer (BL504A, Biosharp) followed by determination of protein concentration using a BCA kit (20201ES76, Yeasen). Subsequently, 20 μg protein samples were separated by SDS-PAGE and transferred to PVDF membranes (IPVH00010, Millipore). The membranes were incubated in 5% non-fat milk at room temperature for 1.5 hours, and then incubated with primary antibody overnight at 4°C, followed by the second antibody for 1 hour at room temperature. The ECL chemiluminescent substrate (BL523A, Biosharp) was then applied onto the membrane (34). Subsequently, the protein bands were observed in a gel imaging system (ChemiDoc XRS+; Bio-Rad Laboratories, Inc.) and finally analyzed with Image J.

All male BALB/c nude mice (6 weeks old) were purchased from Zhuhai, Guangzhou and housed at the Experimental Animal Centre of Hubei University of Chinese Medicine. The temperature was maintained at 26°C-28°C, with relative humidity at 40%-60%, and a 12-hour light/dark cycle. After 1 week of acclimatization, the nude mice were injected subcutaneously with 5×10⁶ HN6 or HN6 (shAMPK) cells. Treatment was initiated when the tumor volume reached 100 mm³. Mice were randomly assigned to three experimental groups of five mice each: 1) Blank control group, administered 200 µL PBS by gavage; 2) CAP group, administered 10 mg/kg CAP dissolved in 200 µL PBS by gavage; 3) CAP+ shAMPK group, administered 10 mg/kg CAP dissolved in 200 µL PBS by gavage (10). Tumor size was measured every 3 days. At the conclusion of the study, the mice were sacrificed, and the excised tumors were collected and weighed. Livers and kidneys were subjected to fixation, embedding, and sectioning, followed by staining with hematoxylin and eosin (HE). Immunohistochemistry (IHC) was performed to evaluate the expression of p-BECN1(Ser90, AP1254, ABclonal, 1:100), ACSL4 (22401-1-AP, Proteintech, 1:100) and GPX4 (30388-1-AP, Proteintech, 1:100) in xenograft tumor tissues. Furthermore, the levels of MDA, Fe²⁺ and GSH in the tumor were assessed using the corresponding analysis kits mentioned above.

The animal handling procedures employed in this research were conducted in accordance with the Guide for the Care and Use of Laboratory Animals established by the National Institutes of Health and were approved by the Institutional Review Board of Hubei University of Chinese Medicine (HUCMS24406389).

Data analysis was performed using GraphPad Prism software version 9.5. The results are presented as the mean ± SEM (Standard Error of Mean) from three independent experiments. Differences between groups were analyzed using one-way ANOVA, and comparisons between two groups were conducted using the t-test. P < 0.05 was considered statistically significant.

HN6 and CAL27 cells were initially pre-treated with 1 μM ferrostatin-1 (Fer-1), a ferroptosis inhibitor. Subsequently, the cells were exposed to 150 μM CAP for 24 hours. CCK-8 assay revealed that CAP significantly inhibited the cell viability in both HN6 and CAL27 cells, while the pre-treatment with Fer-1 partially reversed the effect of CAP ( Figures 1A, B ). These findings suggested that CAP could potentially inhibited the cell proliferation through ferroptosis-dependent pathway.

When ferroptosis occurs, the cells undergoes the accumulation of MDA、Fe²⁺ and ROS, as well as a reduction in GSH levels. Additionally, ferroptosis leads to alterations in mitochondrial structure. Here, TEM results revealed that CAP caused mitochondrial shrinkage, membrane rupture, and cristae fragmentation in both cell lines ( Figure 2A ), which were consistent with the characteristics of ferroptosis in mitochondria (35). The level of lipid oxidation product MDA and Fe²⁺, which served as markers of ferroptosis, was found to be elevated by CAP treatment, while Fer-1 reversed these CAP-induced accumulation ( Figures 2B, C ). Conversely, CAP induced a significant reduction in GSH levels, which could be reversed by treatment with Fer-1 ( Figure 2D ). Following treatment with CAP, fluorescence microscopy results revealed an accumulation of ROS in HN6 and CAL27 cells ( Figure 2E ). Furthermore, GPX4, SLC7A11 and ACSL4 play crucial roles in regulating ferroptosis. Western blot analysis revealed that CAP led to a decrease in GPX4 protein levels and an increase in ACSL4 protein levels in both HN6 and CAL27 cells, though no significant effect was detected on the expression of SLC7A11 ( Figures 2F, G ). The results above provided credible evidences of ferroptosis induced by CAP in TSCC cells.

A previous study has demonstrated that AMPK can phosphorylate BECN1, leading to the formation of the BECN1-SLC7A11 complex, which inhibits SLC7A11 activity and promotes ferroptosis (25). To investigate the reason why the expression of SLC7A11 was unchanged after CAP treatment in HN6 and CAL27 cells, the AMPK/BECN1 signaling pathway was assessed and the results indicated that AMPK was significantly activated and the levels of phosphorylated BECN1 was increased remarkably by CAP ( Figures 3A, B ), which suggested the activation of AMPK/BECN1 signaling pathway. Furthermore, co-immunoprecipitation results revealed that CAP promoted the binding of BECN1 and SLC7A11 in HN6 and CAL27 cells ( Figure 3C ). Meanwhile, the activity of SLC7A11 was suppressed by CAP which was evidenced by a reduction in glutamate release levels ( Figure 3D ). These results suggested that the activity of SLC7A11 was suppressed through AMPK/BECN1 pathway during the ferroptosis induced by CAP in TSCC cells.

ACSL4 is known to trigger ferroptosis within lipid metabolism pathways, whereas GPX4 serves to suppress ferroptosis in antioxidant metabolism pathways. To further explore whether AMPK regulates BECN1 and other molecules, such as ACSL4 or GPX4, in the context of CAP-induced ferroptosis, we constructed a stable cell line with AMPK knockdown through lentivirus-mediated shRNA infection ( Figure 4A ). The results indicated that the knockdown of AMPK could reverse the promoting effect of CAP on p-BECN1 and ACSL4 expression, while having no significant effect on GPX4 expression ( Figures 4B, C ). Meanwhile, the heightened release of glutamate following AMPK knockdown indicated an upregulation of SLC7A11 activity ( Figure 4D ). Furthermore, the accumulation of MDA, Fe²⁺, and ROS, along with the reduction of GSH following AMPK knockdown, reinforced that ferroptosis induced by CAP is regulated not only through the AMPK/BECN1 pathway but also via the AMPK/ACSL4 pathway ( Figures 4E–G ).

In vitro experiments have demonstrated that CAP can induce ferroptosis in TSCC cells through the activation of AMPK. We further conducted investigations to determine whether CAP has similar effects in vivo. Compared to the control group, CAP significantly reduced tumor volume and weight. Additionally, the tumor volume and weight in the CAP + shAMPK group were notably larger than those in the CAP group, indicating that CAP induced ferroptosis in vivo via AMPK ( Figures 5A–C ). There was no significant difference in body weight between the control and treatment groups ( Figure 5D ). No pathological morphological or structural abnormalities were observed in the liver and kidney tissues by HE staining ( Figure 5E ). These findings indicated that CAP has no obviously toxicity to the liver and kidneys of mice. The immunohistochemistry (IHC) results demonstrated that the expression of p-BECN1 and ACSL4 were significantly elevated in the CAP group, while markedly reduced in the group inoculated with shAMPK cells. Nevertheless, the level of GPX4 showed a significant decrease in the CAP group, while no notable recovery was observed in the group inoculated with shAMPK cells ( Figure 5F ). Additionally, elevated levels of MDA and Fe²⁺, along with reduced levels of GSH, were observed in the CAP group, however, these effects were completely reversed in AMPK knockdown group ( Figures 5G–I ), indicating that CAP significantly promotes ferroptosis through AMPK in vivo.