Garcinoic acid promotes ferroptosis-associated growth inhibition in esophageal squamous cell carcinoma through an MDM2-associated mechanism

Introduction

Globally, esophageal cancer (EC) ranks as the seventh leading cause of cancer-related mortality, with approximately 511,000 new cases identified in 2024 (1,2). Histopathological evaluation primarily distinguishes two main subtypes: esophageal squamous cell carcinoma (ESCC) (3) and esophageal adenocarcinoma (EAC) (4), of which ESCC constitutes the predominant form. Recognized risk factors include tobacco use, alcohol consumption, and obesity (5). Because early symptoms are often nonspecific, many patients with ESCC are diagnosed at an advanced stage, when curative surgery is no longer feasible (6). Although concurrent chemoradiotherapy and systemic therapies are available, outcomes for unresectable ESCC remain unsatisfactory, highlighting the need for new therapeutic strategies (1). Hence, developing new chemical entities is imperative to ameliorate the high mortality associated with this malignancy.

Garcinoic acid (GA), a natural analog of delta-tocotrienol (δ-T3), originates from Garcinia kola seeds or is produced by the gut microbiota (7). This compound exhibits significant radical-scavenging, antioxidant, and anti-inflammatory activities (8). These properties are evidenced by its inhibition of 5-lipoxygenase (5-LOX) and its ability to suppress severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)-associated cytokine storms via downregulation of pro-inflammatory cytokines (9). Furthermore, GA functions as an agonist for nuclear receptors such as pregnane X receptor (PXR) and peroxisome proliferator-activated receptor γ (PPAR-γ), demonstrating promise in slowing Alzheimer’s disease progression (10). The anti-inflammatory efficacy of semi-synthetically derived GA is also linked to its inhibition of microsomal prostaglandin E synthase-1 (mPGES-1) (11). Despite these broad pharmacological activities, its specific mechanism of action against ESCC has not been elucidated.

Ferroptosis is a modality of regulated cell death defined by its iron dependence and the pervasive peroxidation of membrane lipids (12,13). The core biochemical drivers of this process encompass intracellular iron overload, GSH depletion, and the consequent accumulation of lipid-reactive oxygen species (ROS) (14). The inability to detoxify these lipid peroxides ultimately precipitates oxidative cell death (15). Morphologically, cells undergoing ferroptosis display characteristic ultrastructural alterations, most notably shrunken mitochondria with increased membrane density and loss of cristae (16). Given this mechanistic understanding, specific small molecules have been developed to precisely induce or suppress ferroptosis by targeting its regulatory pathways. Because mouse double minute 2 (MDM2) is involved in tumor cell survival and may intersect with ferroptosis-regulatory pathways, it represents a plausible candidate mediator linking GA activity to ferroptosis-associated phenotypes in ESCC (17). Its oncogenic functions include negatively regulating the tumor suppressor p53 and being closely linked to aggressive disease features like poor patient prognosis, frequent recurrence, and therapy resistance (18). Whether GA modulates MDM2-related pathways and thereby contributes to ferroptosis-associated growth inhibition in ESCC remains unclear.

This investigation sought to identify the molecular mechanisms underlying the anti-tumor activity of GA in ESCC, with a specific focus on ferroptosis. A comprehensive strategy incorporating network pharmacology and Mendelian randomization (MR) was first utilized to predict GA-associated targets and pathways, followed by experimental validation. The efficacy of GA in suppressing ESCC cell proliferation was then quantitatively determined using in vitro assays. Ultimately, our findings are intended to clarify how GA mediates its therapeutic effect against ESCC, offering valuable insights into future drug development. We present this article in accordance with the MDAR reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2729/rc).

Methods

Prediction of GA protein targets

The initial step involved acquiring the molecular descriptors of GA. Its two-dimensional (2D) chemical structure and corresponding canonical SMILES notation were downloaded from the PubChem database. These descriptors were then utilized for complementary in silico target fishing. The PharmMapper server was queried with the 2D structure file (in SDF format) to generate a set of potential protein targets. Concurrently, the canonical SMILES string was submitted to the SwissTargetPrediction platform to obtain a separate list of putative targets through a ligand-based approach (19). This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Identification of ESCC-associated and ferroptosis-related genes

To establish an ESCC-associated gene set, the GeneCards database was searched using “esophageal squamous cell carcinoma” as the query term (20). Candidate genes were ranked according to their GeneCards relevance scores, and the top 8,000 entries were retained for subsequent analysis. Ferroptosis-related genes were obtained from FerrDb, followed by removal of duplicate records. After curation, 933 unique ferroptosis-related genes were included as the final ferroptosis gene set for downstream intersection and network analyses.

Acquisition of overlapping targets

The list of GA-predicted targets, ferroptosis-related gene and the ESCC-associated target set were first independently curated. The common targets between these two lists were then determined by performing an intersection analysis using the Venny online tool.

Construction of the protein-protein interaction (PPI) network

The shared targets were used to construct a PPI network via the STRING database, with the search parameters restricted to Homo sapiens and a high-confidence minimum interaction score of 0.7. Isolated nodes were excluded from the initial network (21). This network was then imported into Cytoscape for further topological analysis. The five core nodes with the highest degree of connectivity were selected from this network as candidate targets for molecular docking validation.

Molecular docking analysis

The molecular structure of GA was acquired from PubChem and prepared by energy minimization in Chem3D. Concurrently, the crystal structures of the five shortlisted protein targets were downloaded from the RCSB PDB. These protein structures were prepared by removing water molecules and heteroatoms in PyMOL. Both the ligand and receptor files were then converted to the PDBQT format, incorporating necessary hydrogen atoms and partial charges. Docking simulations were executed in AutoDock Vina after defining the binding site grid for each protein. The resulting poses, ranked by predicted binding affinity, were visualized in PyMOL to examine the key intermolecular interactions.

Molecular dynamics simulations

To assess the stability of the GA-MDM2 complex identified by docking, molecular dynamics simulations were run using AMBER 22 (22). The protein was modeled with the ff14SB force field, while GA parameters were derived from GAFF. The system underwent energy minimization in two phases: 2,500 steps of steepest descent followed by 2,500 steps of conjugate gradient. Three independent 100 ns production simulations were then performed. The simulation employed a 2 fs time step, with the Particle Mesh Ewald method for electrostatics and the SHAKE algorithm to constrain bond lengths. Trajectory frames were recorded every 10 ps for analysis.

Cell lines and culture

Human ESCC cell lines EC-109 (Cat. BNCC342591) and EC-9706 (Cat. BNCC339892) were obtained from BeNa Culture Collection (Shanghai, China). The cells were authenticated by STR profiling on May 30, 2022, and the authentication certificates are provided in Figure S1. EC-109 cells were cultured in RPMI-1640 medium, whereas EC-9706 cells were cultured in high-glucose Dulbecco’s Modified Eagle Medium (DMEM). Both media were supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. Cells were maintained at 37 °C in a humidified incubator containing 5% CO₂ and used between passages 4 and 8. Both cell lines were tested using a mycoplasma detection kit (Servicebio, Wuhan, China) before the experiments and were confirmed to be mycoplasma-free.

Preparation of GA and vehicle control

GA was dissolved in dimethyl sulfoxide (DMSO) to prepare a 10 mM GA stock solution and stored at −80 °C. Before each experiment, the stock solution was diluted in complete culture medium to the indicated working concentrations. The final vehicle concentration was maintained at <0.1% in all treatment groups, including the vehicle-only control. The GA concentrations used in this study were 0.3, 0.6, and 1.2 µg/mL, corresponding to 0.70, 1.41, and 2.81 µM, respectively, based on the molecular weight of GA, 426.6 g/mol. The working concentration of 0.6 µg/mL was selected because it produced a reproducible growth-inhibitory effect while avoiding the highest tested concentration; no visible precipitation was observed by light microscopy before cell treatment.

MDM2 overexpression

ESCC cells were transfected with MDM2 overexpression plasmid or the corresponding empty vector (Miaoling Plasmid, Hubei, China) using Lipofectamine 3000 (Servicebio, Wuhan, China) according to the manufacturer’s instructions (16). After 48 h, transfection efficiency was verified by real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR). Cells were then treated with GA at 0.6 µg/mL for 48 h before functional and ferroptosis-associated assays.

Cell counting kit-8 (CCK-8) viability assay

To evaluate cell viability, EC-109 and EC-9706 cells in the logarithmic phase were harvested and seeded into 96-well plates at a density of 5,000 cells per well. After 24 h of adhesion, cells were treated with GA at 0, 0.3, 0.6, and 1.2 µg/mL for 24, 48, and 72 h. Based on the molecular weight of GA, 426.6 g/mol, these concentrations corresponded to 0, 0.70, 1.41, and 2.81 µM, respectively. Vehicle-only controls containing the same final concentration of DMSO were included in parallel. Following the treatment period, a CCK-8 (Servicebio, Wuhan, China) solution was added to each well, and the plates were incubated for 1.5 h. The absorbance at 450 nm was then recorded using a microplate reader to quantify viable cells.

5-Ethynyl-2'-deoxyuridine (EdU) proliferation assay

Cell proliferation was further analyzed with an EdU assay kit (Servicebio, Wuhan, China). Treated cells were incubated with EdU labeling solution, fixed, and permeabilized. A fluorescent Click-iT reaction was performed to detect incorporated EdU. Cell nuclei were visualized by counterstaining with DAPI. Fluorescent images were acquired with an inverted fluorescence microscope following the application of an anti-fade mounting medium.

Acridine orange/ethidium bromide (AO/EB) double staining

The induction of apoptosis in EC-109 and EC-9706 cells was evaluated via AO/EB double staining, employing a commercial assay kit according to the provided instructions. Briefly, cells at 60–70% confluence were exposed to GA at 0.6 µg/mL, equivalent to 1.41 µM, for 48 h, with a vehicle-only control included in parallel. After treatment, the cells were washed, harvested by trypsinization, and resuspended. The cell suspension was subsequently stained with the AO/EB dye mixture and promptly visualized under a fluorescence microscope to distinguish viable, apoptotic, and necrotic cell populations based on their nuclear morphology and staining patterns.

RT-qPCR

Gene expression analysis was conducted through RT-qPCR. Total RNA was first isolated from cell cultures, and 1 µg of the resulting RNA was reverse-transcribed into complementary DNA (cDNA). The qPCR reactions were then carried out using a commercial master mix. The relative mRNA expression of the target gene, MDM2, was normalized to the endogenous control glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and determined via the 2^–ΔΔCt calculation method. The specific primer sequences designed for the amplification are provided in the table below.

• MDM2: F, 5'-AGGCAGGGGAGAGTGATACAGATTC-3'; R, 5'-CAGGAAGCCAATTCTCACGAAGGG-3'.
• GAPDH: F, 5'-CCTGGCACCCAGCACAAT-3'; R, 5'-GGGCCGGACTCGTCATAC-3'.

Transmission electron microscopy (TEM)

For ultrastructural analysis, cells were fixed with 2.5% glutaraldehyde and processed through a series of steps including post-fixation with osmium tetroxide, embolic staining with uranyl acetate, and dehydration. The specimens were then embedded in epoxy resin and polymerized to form blocks. Ultrathin sections were cut, mounted, and examined under a transmission electron microscope. Quantitative assessment of mitochondrial density was performed on the captured micrographs using ImageJ software.

Measurement of ferrous ion levels

To quantify intracellular Fe²⁺, cells subjected to various treatments [GA with or without Ferrostatin-1 (Fer-1)] were washed, harvested, and lysed. The cell lysates were centrifuged, and the resulting supernatants were assayed for ferrous ion concentration using a specialized colorimetric kit, following the supplier’s guidelines.

Measurement of malondialdehyde (MDA) level

Lipid peroxidation was assessed by measuring MDA levels. After exposing cells at 60–70% confluence to GA for 48 h, the cellular MDA content was determined with a commercial assay kit. The absorbance of the reaction product was read at 532 nm for quantification.

Glutathione (GSH) assay

Cellular redox status was evaluated by measuring reduced GSH content using a designated kit. Treated cells were collected and washed to obtain cell pellets, which were then resuspended and lysed. The lysates were incubated with the reaction reagents as specified, and the absorbance at 405 nm was recorded. Final GSH concentrations were derived by interpolation from a standard curve.

Lipid ROS analysis

ESCC cells were plated in 24-well plates 24 h before drug exposure. For lipid ROS detection, cells were treated with GA at 0.6 µg/mL, corresponding to 1.41 µM, for 48 h unless otherwise stated. After treatment, the cells were incubated with 5 µM C11-BODIPY 581/591 probe (Servicebio, Wuhan, China) for 30 min under light-protected conditions. The samples were then harvested, rinsed twice with 1× phosphate-buffered saline (PBS), and resuspended in 500 µL of 1× PBS before fluorescence imaging using a fluorescence microscope (Phoenix, Shangrao, China). Lipid peroxidation was assessed according to the fluorescence transition of C11-BODIPY from red to green, which reflects oxidation of the probe by lipid ROS generated during polyunsaturated lipid peroxidation. The lipid ROS signal was evaluated using green fluorescence ratio of oxidized C11-BODIPY-positive cells, based on 3–6 independent biological replicates and 3–5 randomly selected fields per replicate.

Western blotting

Total cellular proteins were isolated from ESCC cells using RIPA lysis buffer (Servicebio, Wuhan, China). Equivalent quantities of protein were separated by 10% SDS-PAGE and subsequently transferred onto polyvinylidene difluoride (PVDF) membranes (Roche, Germany). The membranes were blocked with QuickBlock^TM Blocking Buffer (Beyotime, Shanghai, China) for 30 min, followed by incubation with the designated primary antibodies at 4 °C overnight. After incubation with appropriate secondary antibodies, protein bands were detected using an enhanced chemiluminescence substrate (Millipore, USA). The primary antibodies used were as follows: anti-GPX4 (GB154327), anti-SLC7A11 (GB155276), anti-ACSL4 (GB155608), and anti-β-actin (GB15003) from Servicebio (Wuhan, China), and anti-MDM2 (CPA1733) from Cohesion Biosciences (Suzhou, China).

MR analysis

A two-sample MR analysis was performed to evaluate whether genetically predicted expression of candidate genes was associated with EC risk (23). Summary-level eQTL data for candidate genes were obtained from the deCODE genetics eQTL summary statistics database and used as the exposure. GWAS summary statistics for EC were used as the outcome. If the outcome dataset was confirmed to be ESCC-specific, it was described as ESCC; otherwise, it was reported as EC.

Single-nucleotide polymorphisms (SNPs) associated with candidate gene expression at genome-wide significance were selected using P<5×10⁻⁸. To reduce linkage disequilibrium, SNPs were clumped using a 10,000-kb window and r²<0.001. Instrument strength was assessed using F-statistics calculated as beta²/SE², and SNPs with F-statistics ≤10 were excluded. For MDM2, three independent SNPs were retained after filtering and clumping, including rs10922098, rs34927613, and rs550671, with F-statistics ranging from 38.68 to 96.80.

Data harmonization and MR analyses were conducted using the TwoSampleMR package in R. The inverse-variance weighted method was used as the primary analysis. Weighted median, weighted mode, simple mode, and MR-Egger regression were used as complementary sensitivity analyses. Cochran’s Q statistics were used to assess heterogeneity, and the MR-Egger intercept was used to evaluate directional horizontal pleiotropy. Leave-one-out analysis was performed to assess the influence of individual SNPs. Effect estimates were reported as beta coefficients and odds ratios (ORs) with 95% confidence interval (CIs). For multiple candidate-gene analyses, false discovery rate (FDR) was applied. MR findings were interpreted as exploratory genetic evidence for candidate-gene prioritization rather than definitive causal proof.

Transwell migration assay

Cell migration was assessed using Transwell chambers. After serum starvation for 24 h, EC-109 and EC-9706 cells were harvested and seeded into the upper chambers at a density of 2×10⁵ cells/mL in serum-free medium. The lower chambers were filled with complete medium containing 10% FBS as a chemoattractant, while the upper chambers contained medium with the specified GA concentrations. Following 24 h of incubation, non-migrated cells on the upper membrane surface were removed with a cotton swab. Cells that had traversed the membrane were fixed with 4% paraformaldehyde, stained with crystal violet, and imaged. The number of migrated cells was determined by counting five randomly selected visual fields per replicate under a light microscope.

Statistical analysis

Data analysis was carried out with GraphPad Prism (v9.0). Continuous variables are expressed as mean ± standard deviation. Unless otherwise specified, each experiment included 3–6 independent biological replicates, with 3–5 technical replicates per biological replicate. For the evaluation of group differences, the Student’s t-test was applied to two-group comparisons, and one-way analysis of variance (ANOVA) was utilized for comparisons across three or more groups. In all analyses, statistical significance was defined as a two-tailed P value below 0.05.

Results

GA suppresses viability, proliferation, and migration while inducing apoptosis in ESCC cells

Given the central role of dysregulated cell proliferation, apoptosis, and migration in ESCC, we first assessed the anti-tumor potential of GA. Initial CCK-8 assays measuring cell viability revealed a concentration- and time-dependent inhibitory effect of GA on both EC-109 and EC-9706 cells (Figure 1A). Based on the corrected dose-response analysis, GA inhibited ESCC cell viability in a concentration- and time-dependent manner at 0.3–1.2 µg/mL, corresponding to 0.70–2.81 µM. A working concentration of 0.6 µg/mL, equivalent to 1.41 µM, for 48 h was selected for subsequent experiments because it produced a reproducible inhibitory effect while avoiding the highest tested concentration. Vehicle-only treatment did not significantly affect cell viability compared with the untreated control (P>0.05, Figure 1B). In contrast, GA markedly reduced EdU incorporation, confirming its anti-proliferative activity (Figure 1C,1D). Furthermore, GA exposure induced significant apoptosis, as evidenced by increased apoptotic rates in AO/EB staining assays (Figure 1E,1F). Finally, Transwell migration assays demonstrated that GA suppressed the migratory capacity of both ESCC cell lines (Figure 1G,1H). Collectively, these findings indicate that GA exerts multifaceted anti-tumor effects against ESCC.

Figure 1 GA inhibits cell viability, proliferation, and migration and induces apoptosis in EC-109 and EC-9706 cells. (A) CCK-8 assay showing ESCC cell viability after treatment with vehicle or GA at 0.3, 0.6, and 1.2 μg/mL for 24, 48, and 72 h. Vehicle-only controls contained <0.1% DMSO and are shown as the DMSO group. (B) Cell viability after treatment with GA at 0.6 μg/mL for 48 h. (C,D) EdU staining showing cell proliferation after GA treatment. (E,F) AO/EB staining showing apoptosis after GA treatment. Scale bars for AO/EB and EdU: 50 μm. (G,H) Transwell migration assays showing migrated cells after GA treatment (crystal violet staining; magnification: ×100). Data are presented as mean ± standard deviation from 3–6 independent biological replicates. ns, P>0.05; *, P<0.05; **, P<0.01; ***, P<0.001, compared with Control group. AO/EB, acridine orange/ethidium bromide; CCK-8, cell counting kit-8; DMSO, dimethyl sulfoxide; EdU, 5-ethynyl-2'-deoxyuridine; ESCC, esophageal squamous cell carcinoma; GA, garcinoic acid; ns, not significant.

GA affects ferroptosis in ESCC cells

To elucidate the mechanism by which GA inhibits ESCC progression, we assessed key hallmarks of ferroptosis. Biochemical analysis demonstrated that GA treatment significantly elevated intracellular iron concentrations (Figure 2A) and concurrently depleted the antioxidant GSH (Figure 2B). Furthermore, a marked increase in MDA, an end-product of lipid peroxidation, was observed (Figure 2C). These biochemical changes are consistent with ferroptosis-associated alterations. TEM further showed mitochondrial morphological changes compatible with ferroptosis, including loss of cristae and increased membrane density (Figure 2D).

Figure 2 GA induces ferroptosis-associated biochemical and mitochondrial changes in ESCC cells. (A-C) Levels of Fe²⁺ (A), GSH (B), and MDA (C) in GA treatment group and control group in EC-109 and EC-9706 cells. (D) GA-treated ESCC cells were processed for TEM analysis. Scale bars: 2 μm. Data are presented as mean ± standard deviation from 3–6 independent biological replicates. **, P<0.01; ***, P<0.001, compared with the control group. ESCC, esophageal squamous cell carcinoma; GA, garcinoic acid; GSH, glutathione; MDA, malondialdehyde; TEM, transmission electron microscopy.

Fer-1 Attenuates GA-Induced Ferroptosis-Associated Phenotypes in ESCC Cells

To ascertain the contribution of ferroptosis to the anti-tumor activity of GA, we co-administered the specific ferroptosis inhibitor Fer-1 with GA and assessed key cellular phenotypes. As illustrated in Figure 3A, Fer-1 significantly alleviated the GA-induced suppression of cell viability. This rescuing effect was corroborated in EdU assays, where Fer-1 co-treatment markedly restored the proliferation capacity attenuated by GA alone (Figure 3B,3C). Similarly, the increased apoptosis rate resulting from GA exposure was substantially reduced upon the addition of Fer-1 (Figure 3D,3E). Finally, the inhibitory effect of GA on cellular migration was also partially, but significantly, reversed by Fer-1 in Transwell assays (Figure 3F,3G). C11-BODIPY staining further showed that GA increased lipid peroxidation, whereas Fer-1 reduced GA-induced lipid ROS accumulation (Figure 3H,3I). Consistently, GA decreased GPX4 and SLC7A11 protein levels and increased ACSL4 protein levels, and these changes were partially attenuated by Fer-1 (Figure 3J,3K). Together, these findings support the involvement of ferroptosis-associated processes in GA-induced ESCC cell death.

Figure 3 Functional rescue experiment after co-intervention of GA and Fer-1. ESCC cells were treated with GA at 0.6 μg/mL for 48 h with or without Fer-1 pretreatment at 1 μM for 12 h. (A) CCK-8 assay. (B,C) EdU staining. (D,E) AO/EB staining. Scale bars for AO/EB and EdU: 50 μm. (F,G) Transwell migration assays were performed following treated GA with or without Fer-1 in EC-109 and EC-9706 (crystal violet staining; magnification: ×100). (H,I) C11-BODIPY staining was quantified as green fluorescence ratio of oxidized based on 3–6 independent biological replicates and 3–5 randomly selected fields per replicate. Red fluorescence indicates non-oxidized C11-BODIPY, whereas green fluorescence indicates oxidized C11-BODIPY. Scale bars: 50 μm. (J,K) Western blot analysis of ferroptosis markers (GPX4, SLC7A11 and ACSL4). Data are presented as mean ± standard deviation from 3–6 independent biological replicates. ^ns, P>0.05; *, P<0.05; **, P<0.01; ***, P<0.001, compared with the control group. ACSL4, acyl-CoA synthetase long-chain family member 4; AO/EB, acridine orange/ethidium bromide; CCK-8, cell counting kit-8; EdU, 5-ethynyl-2'-deoxyuridine; ESCC, esophageal squamous cell carcinoma; Fer-1, Ferrostatin-1; GA, garcinoic acid; GPX4, glutathione peroxidase 4; ns, not significant; SLC7A11, solute carrier family 7 member 11.

Exploratory MR analysis prioritizes MDM2 as a candidate mediator for GA anti-ESCC

The intersection of drug targets and disease-related genes was investigated through a Venn diagram, which identified 24 common targets shared by GA, ferroptosis, and ESCC (Figure 4A). A PPI network constructed from these targets revealed their functional relationships (Figure 4B). To further prioritize candidate genes linking GA-associated targets, ferroptosis, and ESCC, we performed a two-sample MR analysis using eQTL-derived genetic instruments for MDM2 (Figure 4C).

Figure 4 Network pharmacology and exploratory MR analysis prioritize MDM2 as a candidate mediator. (A) Venn diagram showing overlapping targets among GA-associated targets, ferroptosis-related genes, and ESCC-associated genes. (B) Protein-protein interaction network of overlapping genes. Node size and color intensity represent degree values. (C) Primary IVW estimate for the association between genetically predicted MDM2 expression and ESCC risk. (D) Scatter plot showing MR estimates from different MR methods. (E) Leave-one-out sensitivity analysis for MDM2-related genetic instruments. (F) Representative MDM2 expression data in normal and tumor tissues from HPA database. Images were obtained from the HPA database and are shown as representative database-derived evidence. IHC staining of MDM2 in normal esophageal tissue (https://www.proteinatlas.org/ENSG00000135679-MDM2/tissue/esophagus) and head and neck cancer tissue (https://www.proteinatlas.org/ENSG00000135679-MDM2/cancer/head+and+neck+cancer) from the HPA database. ESCC, esophageal squamous cell carcinoma; GA, garcinoic acid; HPA, Human Protein Atlas; IHC, immunohistochemical; IVW, inverse-variance weighted; MDM2, mouse double minute 2; MR, Mendelian randomization.

In the primary inverse-variance weighted analysis, genetically predicted MDM2 expression showed a nominal positive association with ESCC risk [OR =1.00193; 95% CI: 1.00005–1.00382; P_inverse-variance weighted (IVW) =0.044]. Complementary MR methods showed directionally consistent but statistically weaker estimates (Figure 4D). Sensitivity analyses did not provide evidence of substantial heterogeneity or directional horizontal pleiotropy. Cochran’s Q test showed no significant heterogeneity in either the IVW model (Q =0.625, P=0.73) or the MR-Egger model (Q =0.537, P=0.46). The MR-Egger intercept was close to zero and not statistically significant (intercept =0.000108, P=0.81), suggesting no clear evidence of directional pleiotropy among the included instruments. Leave-one-out analysis showed that the direction of the IVW estimate remained positive after removing each SNP individually; however, the nominal statistical significance was attenuated after excluding rs34927613 or rs550671. Therefore, the MR result was interpreted as exploratory genetic evidence supporting the prioritization of MDM2, rather than definitive proof of a causal relationship (Figure 4E).

Because the current analysis retained only three genetic instruments and the complementary MR methods were not uniformly significant, we avoided overinterpreting the MR findings. Together with the network pharmacology and subsequent experimental validation, these results supported MDM2 as a candidate mediator for further mechanistic investigation in the context of GA-associated ferroptosis and growth inhibition. Consistent with the prioritization of MDM2, database-derived representative images from the Human Protein Atlas suggested higher MDM2 staining intensity in head and neck cancer tissue than in normal esophageal tissue (Figure 4F). These database-derived images were used as descriptive supporting evidence and were not interpreted as genetic or causal evidence.

In silico analyses predict a potential interaction between GA and MDM2

Molecular docking was employed to predict the binding mode and affinity between GA and the MDM2 protein. The results indicated a predicted favorable interaction, with a calculated binding energy of −9.1 kcal/mol (Figure 5A). This binding is primarily stabilized by hydrogen bonds formed between GA and key residues (His-96 and Lys-51) within the MDM2 binding pocket. To evaluate this docked complex, we performed a 100 ns molecular dynamics simulation. The root-mean-square deviation (RMSD) of the protein backbone equilibrated early in the simulation and remained stable, suggesting the system reached a converged state (Figure 5B). Analysis of root-mean-square fluctuation (RMSF) provided insights into regional residue flexibility (Figure 5C). The persistence of intermolecular hydrogen bonds throughout the simulation trajectory further corroborated a stable binding mode (Figure 5D). Additionally, the solvent-accessible surface area (SASA) and the radius of gyration (Rg) of the complex remained largely constant (Figure 5E-5G), suggesting the structural integrity and compactness of the GA-MDM2 complex under simulated physiological conditions.

Figure 5 In silico prediction of the GA-MDM2 interaction. (A) 3D and 2D interaction diagrams of the GA and MDM2 complex. (B) RMSD of the GA and MDM2 complex. (C) RMSF of the GA and MDM2 complex. (D) Hydrogen bonds of GA and MDM2 complex. (E) SASA of GA and MDM2 complex. (F) Rg of the GA and MDM2 complex. (G) Free-energy landscape of the GA-MDM2 complex based on RMSD and Rg. 2D, two-dimensional; 3D, three-dimensional; GA, garcinoic acid; MDM2, mouse double minute 2; Rg, radius of gyration; RMSD, root-mean-square deviation; RMSF, root-mean-square fluctuation; SASA, solvent-accessible surface area.

MDM2 overexpression attenuates GA-induced ferroptosis-associated phenotypes

To examine the functional relevance of MDM2 in GA-induced phenotypes, ESCC cells were transfected with an MDM2 overexpression construct, and transfection efficiency was confirmed by RT-qPCR (Figure S2). Western blotting showed that GA reduced MDM2 protein abundance, whereas MDM2 overexpression increased MDM2 protein levels under GA treatment (Figure 6A). Functionally, MDM2 overexpression attenuated GA-induced reductions in cell viability and proliferation and reduced GA-induced apoptosis (Figure 6B-6D). MDM2 overexpression also partially restored cell migration and mitigated GA-induced ferroptosis-associated changes, including Fe²⁺ accumulation, MDA elevation, GSH depletion, lipid ROS accumulation, and mitochondrial morphological alterations (Figure 6E-6J). These findings support the functional relevance of MDM2 in GA-induced ferroptosis-associated phenotypes.

Figure 6 MDM2 overexpression attenuates GA-induced ferroptosis-associated phenotypes in ESCC cells. (A) MDM2 protein expression was assessed by Western blotting. (B,C) Cell viability and EdU incorporation after GA treatment with or without MDM2 overexpression. (D) AO/EB staining showing apoptosis after GA treatment with or without MDM2 overexpression. (E) Transwell migration assay showing migrated cells. (F-H) Intracellular Fe²⁺ (F), MDA (G), and GSH (H) levels in EC-109 and EC-9706 cells. (I) C11-BODIPY staining showing lipid ROS accumulation. (J) TEM images showing mitochondrial morphology. Scale bars:. Data are presented as mean ± standard deviation from 3–6 independent biological replicates. *, P<0.05; **, P<0.01; ***, P<0.001, compared with the control group or GA group. AO/EB, acridine orange/ethidium bromide; EdU, 5-ethynyl-2'-deoxyuridine; ESCC, esophageal squamous cell carcinoma; GA, garcinoic acid; GSH, glutathione; MDA, malondialdehyde; MDM2, mouse double minute 2; ROS, reactive oxygen species; TEM, transmission electron microscopy.

Discussion

The clinical management of ESCC continues to be hampered by significant challenges, including dose-limiting toxicities and the development of resistance to conventional therapies, which collectively constrain treatment outcomes (24,25). This underscores the demand for new therapeutic entities with superior efficacy and tolerability. GA has previously been characterized by its anti-inflammatory and antioxidant properties, suggesting broader pharmacological potential (26). Nonetheless, its capacity to modulate ESCC pathogenesis and the specific mechanisms involved remained largely unexplored. In this study, we found that GA suppressed ESCC cell viability, proliferation, and migration while promoting cell death in vitro. Mechanistically, our data support the involvement of ferroptosis-associated changes and suggest that MDM2 is a candidate mediator of GA-induced phenotypes. The combined use of network pharmacology, exploratory MR analysis, in silico modeling, and functional rescue assays provides a framework for further investigating the GA-MDM2-ferroptosis relationship in ESCC.

Ferroptosis represents an iron-dependent form of regulated cell death characterized by the pervasive peroxidation of membrane lipids, driven by metabolic imbalances in iron and ROS homeostasis (27,28). Given its defined biochemical basis, this cell death modality presents an attractive target for therapeutic intervention in ESCC (29,30). In the present study, we observed that GA treatment elicits the ferroptosis-associated signature in ESCC cells, manifesting as intracellular iron and lipid peroxide accumulation, GSH depletion, and mitochondrial membrane depolarization. The critical involvement of ferroptosis was further substantiated by experiments showing that the ferroptosis inhibitor Fer-1 significantly attenuated GA-induced cell death and partially attenuated the associated biochemical changes. Complementing these findings, ultrastructural analysis via TEM provided visual evidence of ferroptosis mitochondria, exhibiting the characteristic phenotype of reduced volume, disintegrated cristae, and enhanced membrane density. Collectively, these findings should be interpreted as evidence for ferroptosis-associated cell death rather than proof that ferroptosis is the exclusive mode of GA-induced cell death.

Our integrated bioinformatics analysis positioned MDM2 as a critical node connecting GA’s pharmacological activity with ferroptosis and ESCC pathogenesis. MDM2 is a well-established oncogenic regulator (17) and is best known for modulating p53-dependent pathways (32). Although our data support the functional relevance of MDM2 in GA-induced ferroptosis-associated phenotypes, we did not directly assess p53 activity, MDM2-p53 interaction, or direct GA-MDM2 target engagement. Therefore, future studies are needed to determine whether the MDM2-p53 axis or other MDM2-related pathways contribute to the observed effects. Supporting this premise, MDM2 is known to be overexpressed in ESCC (33), and our computational models predicted a stable binding interaction between GA and MDM2. Functional validation using MDM2-overexpressing cells demonstrated that elevated MDM2 levels significantly counteract the anti-proliferative and pro-ferroptotic effects of GA. Functional validation using MDM2-overexpressing cells showed that elevated MDM2 levels attenuated GA-induced anti-proliferative and ferroptosis-associated phenotypes. These findings support the functional relevance of MDM2 in this context, but additional loss-of-function and target-engagement experiments are required to define whether MDM2 is necessary for GA-induced ferroptosis-associated effects.

This work is subject to several conceptual and technical limitations. First, although the working concentration of GA was selected based on dose-response analysis, vehicle-control experiments, and the absence of visible precipitation, its selectivity toward malignant versus non-malignant esophageal epithelial cells remains to be further evaluated. Second, the present data support the involvement of ferroptosis in GA-induced ESCC cell death; however, they do not exclude the contribution of other regulated cell death pathways. Specifically, C11-BODIPY oxidation, changes in ferroptosis-associated proteins, and Fer-1 rescue collectively support a ferroptosis-associated mechanism, but additional rescue strategies, such as iron chelation or alternative lipid peroxidation inhibitors, would further strengthen the specificity of this conclusion. Third, molecular docking and molecular dynamics simulations predicted a stable GA-MDM2 interaction, and MDM2 overexpression attenuated GA-induced phenotypes. Nevertheless, these findings do not establish direct target engagement between GA and MDM2. Future studies using biochemical binding assays or cellular target-engagement experiments are therefore needed to determine whether GA directly interacts with MDM2.

Conclusions

Collectively, our findings suggest that GA suppresses ESCC cell growth and migration and promotes ferroptosis-associated changes, at least partly through an MDM2-associated mechanism. Because direct GA-MDM2 target engagement and MDM2 loss-of-function evidence remain to be established, the MDM2-ferroptosis axis should be considered a promising direction for further mechanistic and therapeutic investigation.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the MDAR reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2729/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2729/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2729/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2729/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Jiang W, Zhang B, Xu J, et al. Current status and perspectives of esophageal cancer: a comprehensive review. Cancer Commun (Lond) 2025;45:281-331. [Crossref] [PubMed]
2. Yang H, Wang F, Hallemeier CL, et al. Oesophageal cancer. Lancet 2024;404:1991-2005. [Crossref] [PubMed]
3. Zhan J, Li M, Li L, et al. Targeting LTBP2 Derived from Cancer-Associated Fibroblasts Sensitizes Esophageal Squamous Cell Carcinoma to Chemotherapy. Cancer Res 2025;85:2412-28. [Crossref] [PubMed]
4. Yates J, Mathey-Andrews C, Park J, et al. Cell states and neighborhoods in distinct clinical stages of primary and metastatic esophageal adenocarcinoma. Cell Rep Med 2025;6:102188. [Crossref] [PubMed]
5. Qi L, Wang J, Hou S, et al. Unraveling the tumor microenvironment of esophageal squamous cell carcinoma through single-cell sequencing: A comprehensive review. Biochim Biophys Acta Rev Cancer 2025;1880:189264. [Crossref] [PubMed]
6. Ren W, Zhang H, Li Y, et al. Efficacy and safety of PD-1/PD-L1 inhibitors as first-line treatment for esophageal squamous cell carcinoma: a systematic review and meta-analysis. Front Immunol 2025;16:1563300. [Crossref] [PubMed]
7. Isot I, Kim SH, Demirel-Yalciner T, et al. Garcinoic acid enhances inflammation resolution against colitis by activating Nrf2 dependent efferocytosis. Free Radic Biol Med 2025;237:37-45. [Crossref] [PubMed]
8. Börmel L, Geisler AR, Hupfer Y, et al. The Vitamin E Derivative Garcinoic Acid Suppresses NLRP3 Inflammasome Activation and Pyroptosis in Murine Macrophages. Inflammation 2025;48:3340-52. [Crossref] [PubMed]
9. Olajide OA, Iwuanyanwu VU, Lepiarz-Raba I, et al. Garcinia kola and garcinoic acid suppress SARS-CoV-2 spike glycoprotein S1-induced hyper-inflammation in human PBMCs through inhibition of NF-κB activation. Phytother Res 2021;35:6963-73. [Crossref] [PubMed]
10. Marinelli R, Torquato P, Bartolini D, et al. Garcinoic acid prevents β-amyloid (Aβ) deposition in the mouse brain. J Biol Chem 2020;295:11866-76. [Crossref] [PubMed]
11. Alsabil K, Suor-Cherer S, Koeberle A, et al. Semisynthetic and Natural Garcinoic Acid Isoforms as New mPGES-1 Inhibitors. Planta Med 2016;82:1110-6. [Crossref] [PubMed]
12. Jiang K, Yin X, Zhang Q, et al. STC2 activates PRMT5 to induce radioresistance through DNA damage repair and ferroptosis pathways in esophageal squamous cell carcinoma. Redox Biol 2023;60:102626. [Crossref] [PubMed]
13. Peng H, He Y, Hu Y, et al. Berbamine promotes ferroptosis of esophageal squamous cell carcinoma by facilitating USP51-mediated GPX4 ubiquitination and degradation. Biomed Pharmacother 2024;179:117309. [Crossref] [PubMed]
14. Zhao M, Lu T, Bi G, et al. PLK1 regulating chemoradiotherapy sensitivity of esophageal squamous cell carcinoma through pentose phosphate pathway/ferroptosis. Biomed Pharmacother 2023;168:115711. [Crossref] [PubMed]
15. Mi Y, Chen L, Wang C, et al. AURKA knockdown inhibits esophageal squamous cell carcinoma progression through ferroptosis. Heliyon 2024;10:e28365. [Crossref] [PubMed]
16. Peng H, Yan Y, He M, et al. SLC43A2 and NFκB signaling pathway regulate methionine/cystine restriction-induced ferroptosis in esophageal squamous cell carcinoma via a feedback loop. Cell Death Dis 2023;14:347. [Crossref] [PubMed]
17. Yang A, Song H, Cheng Y. The GNL3L-MDM2 Interaction Drives Esophageal Squamous Cell Carcinoma Progression. Cancer Med 2025;14:e71146. [Crossref] [PubMed]
18. Chai Y, Peng B, Dai L, et al. Autoantibodies response to MDM2 and p53 in the immunodiagnosis of esophageal squamous cell carcinoma. Scand J Immunol 2014;80:362-8. [Crossref] [PubMed]
19. Zhang S, Guo J, Zhang H, et al. Gliotoxin Induced Ferroptosis by Downregulating SUV39H1 Expression in Esophageal Cancer Cells. Recent Pat Anticancer Drug Discov 2023;18:397-407. [Crossref] [PubMed]
20. Yang Z, Chang C, Gao N, et al. Integrating network pharmacology and molecular modeling to decipher the anti-esophageal squamous cell carcinoma mechanisms of Bidens pilosa L. Discov Oncol 2025;16:1589. [Crossref] [PubMed]
21. Shen F, Liu X, Ding F, et al. Integrative bioinformatics analysis of high-throughput sequencing and in vitro functional analysis leads to uncovering key hub genes in esophageal squamous cell carcinoma. Hereditas 2025;162:38. [Crossref] [PubMed]
22. Davar D, Eroglu Z, Milhem M, et al. Combined Targeting of PD-1 and TIM-3 in Patients with Locally Advanced or Metastatic Non-Small Cell Lung Cancer: AMBER Part 2B. Clin Cancer Res 2025;31:3443-51. [Crossref] [PubMed]
23. Amin N, Liu J, Bonnechere B, et al. Interplay of Metabolome and Gut Microbiome in Individuals With Major Depressive Disorder vs Control Individuals. JAMA Psychiatry 2023;80:597-609. [Crossref] [PubMed]
24. Shang X, Xie Y, Yu J, et al. A prospective study of neoadjuvant pembrolizumab plus chemotherapy for resectable esophageal squamous cell carcinoma: The Keystone-001 trial. Cancer Cell 2024;42:1747-1763.e7. [Crossref] [PubMed]
25. Guo X, Chen C, Zhao J, et al. Neoadjuvant Chemoradiotherapy vs Chemoimmunotherapy for Esophageal Squamous Cell Carcinoma. JAMA Surg 2025;160:565-74. [Crossref] [PubMed]
26. Dinh CP, Ville A, Neukirch K, et al. Structure-based design, semi-synthesis and anti-inflammatory activity of tocotrienolic amides as 5-lipoxygenase inhibitors. Eur J Med Chem 2020;202:112518. [Crossref] [PubMed]
27. Song J, Zhang J, Shi Y, et al. Hypoxia inhibits ferritinophagy-mediated ferroptosis in esophageal squamous cell carcinoma via the USP2-NCOA4 axis. Oncogene 2024;43:2000-14. [Crossref] [PubMed]
28. Qiao G, Zhang W, Dong K. Regulation of ferroptosis by noncoding RNAs: a novel promise treatment in esophageal squamous cell carcinoma. Mol Cell Biochem 2022;477:2193-202. [Crossref] [PubMed]
29. Yang R, Wan J, Ma L, et al. TMEM44-AS1 promotes esophageal squamous cell carcinoma progression by regulating the IGF2BP2-GPX4 axis in modulating ferroptosis. Cell Death Discov 2023;9:431. [Crossref] [PubMed]
30. Zheng Y, Liu F, Huang Y, et al. SNORA58 Facilitates Radioresistance via Suppressing JNK1-Mediated Ferroptosis in Esophageal Squamous Cell Carcinoma. Adv Sci (Weinh) 2025;12:e08515. [Crossref] [PubMed]
31. Jensen GF, Brownfield D. Gender, lifestyles, and victimization: beyond routine activity. Violence Vict 1986;1:85-99.
32. Bedolla N, Liu L, Liu X, et al. Ursolic acid enhances radiosensitivity in esophageal squamous cell carcinoma by modulating p53/SLC7A11/GPX4 pathway-mediated ferroptosis. Toxicon 2025;255:108233. [Crossref] [PubMed]
33. Sawada R, Maehara R, Oshikiri T, et al. MDM2 copy number increase: a poor prognostic, molecular event in esophageal squamous cell carcinoma. Hum Pathol 2019;89:1-9. [Crossref] [PubMed]