Platelet-activating factor induces ferroptosis by binding to ATF3 and inhibiting the SLC7A11/GPX4 axis to suppress the progression of endometrial carcinoma Authors: Feifei Pan, Yuqing Qiu, Zhuojie Huang, Yanqin Zheng, Yun Chen, Yiming Song, Zhe-Sheng Chen, Lirong Guo, Jingyao Wang, Xiaojie Liu, Zixin Tao, Yaqiong Liu, Xingcui Xuan, Kunxiang Gong, Kun Shi Published: 2025 ABSTRACT AbstractBackgroundEndometrial carcinoma (EC) is an aggressive gynecological malignancy with high recurrence rates and limited treatment options. Patients with EC exhibit elevated serum levels of lyso-platelet-activating factor (lyso-PAF), a biologically inactive product generated during the hydrolysis of PAF by PAF-acetylhydrolase (II) (PAF-AH2). Nevertheless, the biological function of PAF in EC progression and the relationship between abnormal lyso-PAF levels and EC progression remain unclear.ResultsIn this study, we identified the bioactive phospholipid mediator–PAF, which effectively suppress tumor growth both in vitro and in patient-derived organoid models. By combining the cellular thermal shift assay, drug affinity responsive target stability approach, molecular docking simulation, and surface plasmon resonance analysis, we identified activating transcription factor 3 (ATF3) as the direct binding target of PAF with a Kd of 3 0.24 µM. Mechanistically, PAF enhanced ATF3 stability through deubiquitination. In addition, ATF3 knockdown partially reversed the effects of PAF-induced SLC7A11 and GPX4 transcriptional inhibition. In vivo studies using a nude mouse xenograft model confirmed the ferroptosis-mediated tumor-suppressive effects of PAF, as evidenced by hematoxylin and eosin staining and immunohistochemical analysis. Clinically, proteomic analysis revealed upregulation of PAF-AH2 expression in EC tissues derived from patients. Subsequent analysis of EC tissue microarrays revealed that high PAF-AH2 expression is associated with poor survival.ConclusionsIn conclusion, this study demonstrates that PAF inhibits EC tumor growth by inducing ferroptosis and that the underlying mechanisms of its action involve the targeting of ATF3. These findings suggest that PAF metabolism may be a promising direction for EC diagnostics and therapeutics.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13062-025-00713-z.Keywords: Endometrial carcinoma, Platelet-activating factor, Ferroptosis, ATF3, PAF-AH2 Background Endometrial cancer (EC) is a highly prevalent gynecologic cancer, with 420,242 new cases and 97,704 deaths reported globally in 2022 [1]. Originating from the endometrial epithelium, EC is divided into two distinct histological subtypes: endometrioid carcinoma and non-endometrioid carcinoma. Endometrioid carcinoma accounts for the majority of EC cases and involves the expression of estrogen and progesterone receptors, making it a potential target for hormone-based therapy. Non-endometrioid carcinomas account for approximately 20% of cases and include serous and clear cell carcinomas. In contrast to endometrioid carcinomas, non-endometrioid carcinomas exhibit estrogen-independent growth patterns, demonstrate more aggressive clinical behavior, and are associated with poorer prognosis [2, 3]. The development of EC is a complex, multistep process influenced by various micro-environmental factors. However its exact etiology remains unclear. Endogenous metabolites directly affect tumor progression [4]. We have recently demonstrated that the accumulation of phosphocholine (PC) inhibits EC cell proliferation and stem cell characteristics by downregulating the mTOR‑c‑Myc signaling pathway [5]. Moreover, in patients with EC, serum metabolite profiling revealed elevated levels of PC and lysoplatelet-activating factor-16 (lyso-PAF) [6]. Lyso-PAF is a biologically inactive product generated during the conversion of platelet-activating factor (PAF) by PAF-AH2 [7]. PAF is a bioactive phospholipid mediator that regulates diverse biological processes, including allergic reactions, inflammation, and tumor progression [8]. However, the biological functions and mechanisms of PAF in EC progression remain unclear. Ferroptosis is a form of programmed cell death characterized by the iron-dependent accumulation of lethal amounts of reactive oxygen species (ROS), leading to oxidative stress and cell death [9]. It is increasingly recognized as a critical mechanism of cell death induced by chemotherapy and targeted therapies in cancer [10]. Cellular sensitivity to ferroptosis is regulated by key cellular processes, such as polyunsaturated fatty acid (PUFA)-containing phospholipid synthesis and peroxidation, glutathione metabolism, and intracellular iron ion metabolism [11, 12]. Notably, PAF and PAF-like phospholipids initiate and propagate ferroptosis, contributing to renal tubular injury during acute kidney injury [13]. Therefore, we hypothesize that PAF could mediates EC progression by modulating ferroptosis. However, how PAF influences EC progression via pathways such as ferroptosis remains unknown. Our results revealed PAF as a tumor progression suppressor in EC. Moreover, we demonstrated that PAF exerts its tumor-suppressing effect by binding to ATF3. This study highlights PAF metabolism as a basis for future diagnostic and therapeutic strategies in EC. Methods Cell culture and transfection HEC-1B (STCC10605P), AN3CA (STCC10601P), and 293T (STCC10301P) cell lines were purchased from Servicebio (Hubei, China) and were recently authenticated by Short Tandem Repeat profiling. The cells were cultured in DMEM complete medium, containing DMEM basic medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 10% fetal bovine serum (Procell Life Science & Technology Co. Ltd., Wuhan, China), and 1% penicillin-streptomycin (Gibco), at 37 °C with 5% CO2. All cells used in the experiments were in the logarithmic growth phase within 20 passages. Plasmids and siRNAs were transfected using Lipofectamine 3000 (L3000015; Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s protocol. Plasmids and siRNAs were diluted in Opti-MEM. Cell viability assay Cell viability was assessed using the cell-counting kit 8 (CCK-8) assay. Briefly, cells were seeded overnight in 96-well plates (5 × 103 cells/well) and treated with PAF. After incubation for 48 h, the CCK8 assay (C0039; Beyotime, Suzhou, China) was performed according to the manufacturer’s protocol, and the absorbance was recorded at 450 nm. 5-ethynyl-2-deoxyuridine (EdU) incorporation assay. EdU incorporation assay was performed to investigate DNA synthesis using an EdU Cell Proliferation Kit with Alexa Fluor 555 (C0075S; Beyotime) following the manufacturer’s protocol. Briefly, after 24-h treatment with PAF (60 µM), cells were incubated with 50 µM EdU buffer at 37 °C for 2 h, fixed with 4% formaldehyde for 15 min, and permeabilized with 0.1% Triton X-100 for 10 min. Cells were then sequentially stained with Alexa Fluor 555 azide and Hoechst 33,342 and imaged under a confocal microscope (Nikon, Tokyo, Japan). Colony formation assay For the colony formation assay, cells were transferred to 12-well plates (200 cells/well) and treated with PAF or vehicle control for 24 h. Subsequently, the medium was replaced with fresh complete medium, and the culture was continued for 9–12 d. The cells were then mixed with methanol, stained with 0.1% crystal violet, and imaged using a camera. Patient-derived organoid (PDO) culture This study was conducted in accordance with the Declaration of Helsinki and approved by the Medical Ethics Committee of Guangzhou Women and Children’s Medical Center (Approval number: 124B00). Patient-derived EC organoids were developed as described previously [14]. Tumor biopsy samples were transported in ice-cold Advanced DMEM/F12 (12634028; Thermo Fisher Scientific, USA) containing 1% penicillin-streptomycin. The biopsy specimens were minced and digested at 37 °C for 1 h using the Advanced DMEM/F12 solution containing type IV collagenase (17104019; Thermo Fisher Scientific, USA). Following filtration, the samples were centrifuged at 300 × g for 5 min, seeded into Matrigel in a pre-warmed 24-well plate, and cultured in 500 µL of basal growth media. The next day, the organoids were treated with PAF at the indicated concentrations for 2 d. Cell viability of the patient-derived organoids was determined using the Cell-Counting-Lite 3D Luminescent Cell Viability Assay (DD1102-01; Vazyme, China). Xenograft mouse models All animal experiments were approved by the Guangzhou Women and Children’s Medical Center Animal Welfare and Ethics Committee (RSDW­2023­01427) and were conducted in accordance with the NIH Guidelines for the Care and Use of Laboratory Animals. Four-week-old female nude BALB/c mice were maintained in a specific pathogen-free facility, with a constant temperature of 22–25 °C, 45–65% humidity, and a 12-h light–dark cycle. The AN3CA and HEC1B xenograft models were established by subcutaneously (s.c.) injecting 5 × 105 cells suspended in Matrigel (356230; Corning Inc., NY, USA) into the flank of nude mice (Jicui Yaokang Biological Co. Ltd., Nanjing, China). Mice were randomized into two groups, vehicle (saline solution) and PAF (0.25 mg·kg− 1), to receive the treatments via peritumoral injection every 3 days. The dissected xenograft tumor tissues were weighed and fixed in 4% paraformaldehyde for immunohistochemistry (IHC) analysis. The tumor volume was calculated as follows: tumor volume = 0.5 × length × width2. Western blot (WB) and cellular thermal shift assays (CETSA) WB analysis was performed on cells treated with different concentrations of PAF for 48 h. Cells were washed with phosphate-buffered saline (PBS) and lysed in RIPA lysis buffer containing protease and phosphatase inhibitors. Following centrifugation (12000 ×g, 20 min, 4 °C), the protein concentrations of the supernatant were determined using a BCA Protein Assay kit (23227; Thermo Fisher Scientific, USA) according to the manufacturer’s protocol. Equal volume of protein samples were separated by SDS-PAGE, transferred onto PVDF membranes, and then probed using the following antibodies at the indicated dilutions: NRF2 (380773, 1:1000; Zen BioScience, China), ACSL4 (R24265, 1:1000; Zen BioScience, China), SLC7A11 (382036, 1:1000; Zen BioScience, China), FTH1 (381204, 1:1000; Zen BioScience, China), GPX4 (bs-3884R, 1:1000; Biosss, China), ATF3 (18665 S, 1:1000; Cell Signaling Technology, USA), and GAPDH (5174, 1:1000; Cell Signaling Technology, USA). For CETSA, the cell lysates were treated with either PBS or PAF (100 µM) for 3 h at 37 °C with shaking. The samples were then aliquoted into separate tubes and heated at different temperatures for 3 min. After cooling, the lysates were centrifuged at 12,000 × g for 20 min at 4 °C, and the supernatant was collected for subsequent WB analysis. Drug affinity responsive target stability (DARTS) DARTS was performed as previously described [15]. Briefly, cell lysates were aliquoted into 1.5 mL tubes and incubated with PBS or PAF for 3 h at 25 °C. After incubation, the cell lysates were digested with pronase E (HY-114158; MedChemExpress, China) for 30 min at 37 °C. Subsequently, a protein loading buffer was immediately added, and the lysate samples were heated to stop proteolysis. WB analysis was performed for further analysis. RNA isolation and quantitative real-time PCR (qRT-PCR) analysis qRT-PCR analysis was performed as previously described [16]. Briefly, total RNA was isolated from AN3CA and HEB1C cells using an animal total RNA isolation kit (Foregene, Chengdu, China). Total RNA was reverse-transcribed into complementary DNA using the PrimeScript™ RT Reagent Kit (Takara, China) according to the manufacturer’s instructions. The resulting cDNA was used for qPCR analysis using the PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific) on a LightCycler 480 system (Roche, Basel, Switzerland). All data were normalized to Gapdh and analyzed following the 2−ΔΔCt method. The primer sequences were as follows: forward primers: ATF3 (5′-CCTCTGCGCTGGAATCAGTC-3′), DDIT3 (5′-GGAAACAGAGTGGTCATTCCC-3′). Reverse primers: ATF3 (5′-TTCTTTCTCGTCGCCTCTTTTT-3′), DDIT3 (5′-CTGCTTGAGCCGTTCATTCTC-3′). RNA sequencing AN3CA and HEC1B cells were collected after 48 h treatment with PAF (60 µM), and total RNA was extracted with TRIzol (Invitrogen) according to the manufacturer’s instructions. The amount, quality, and composition of the isolated RNA were analyzed using a NanoDrop Spectrophotometer (Thermo Fisher Scientific), and total RNA integrity was analyzed using an Agilent 2100 Bioanalyzer (Thermo Fisher Scientific). Total RNA enrichment for sequencing poly(A) RNAs was performed using a TruSeq mRNA sample preparation kit (Illumina, San Diego, CA, USA) according to the manufacturer’s instructions. After thermal fragmentation at an appropriate temperature, small pieces of the fragmented poly(A) RNA-enriched samples were subjected to cDNA synthesis using an Illumina TruSeq preparation kit. The cDNA products were purified using Ampure XP Beads and validated using an Agilent Technologies 2100 Bioanalyzer for quality control. Subsequently, the products were heat-denatured and circularized using a splint oligo sequence to obtain the final library. This library, prepared as single-stranded circular DNA, was then amplified via phi29 into DNA nanoballs, each containing more than 300 copies of a single molecule, and loaded into patterned nanoarrays to generate paired-end 150-base reads on the MGIseq2000 platform (UW Genetics Shenzhen, China). The RNA sequencing results are presented in Additional file 1. Surface plasmon resonance (SPR) analysis SPR assays were performed by TopScience Co., Ltd. (Shanghai, China) to detect the binding between the recombinant ATF3 protein (TMPH-01164, TargetMol) and C16-PAF (74389-68-7, TargetMol), with the ATF3 protein and anti-ATF3 antibody combination used as the positive control. The ATF3 protein was immobilized on a CM5 sensor chip (BR-1005-30; Cytiva, Marlborough, MA, USA). The anti-ATF3 antibody and PAF were diluted to a series of concentrations in PBS and then allowed to flow at 10 µL min–1 for 150 s in each run. The experiments were conducted on a Biacore T200 instrument (v.2.0; GE HealthCare, Chicago, IL, USA), and the dissociation constant (KD) values were determined based on the 1:1 Langmuir binding model using Biacore T200 Evaluation software (v.2.0; GE Healthcare) following the manufacturer’s guidelines. Ubiquitination assay Flag-ATF3 and HA-ubiquitin plasmids were co-transfected into HEK293 cells using Lipofectamine 3000 reagent (Lipo3000; Invitrogen) for 48 h. The cells were then treated with PAF (100 µM) for 3 h and subjected to pull-down assays using an anti-DYKDDDDK-tag FAST IP kit (EA-IP-K001; Elabscience, China). Cell lysates were immunoprecipitated with anti-FLAG to pull down ATF3 and immunoblotted for HA-ubiquitin to assess ATF3 ubiquitination after treatment with PAF. ROS assay ROS levels in AN3CA and HEC1B cells were evaluated using a ROS assay kit (E-BC-K138; Elabscience). Briefly, after the indicated treatment, AN3CA and HEC1B cells were treated with 10 µM DCFH-DA (1:1000) and incubated at 37 °C in the dark for 30 min. The cells were washed twice with PBS, trypsinized, and their fluorescence was measured using a flow cytometer (BD Biosciences, San Jose, CA, USA). Ferrous iron content in cells The levels of ferrous iron in the cells were measured using a Cell Ferrous Iron Colorimetric Assay Kit (E-BC-K881-M; Elabscience). Cells were seeded in 6-well plates (1 × 106 cells/well) and treated with PAF for 48 h, followed by trypsinization and collection. The concentration of ferrous iron was determined according to a protocol provided by the manufacturer. The luminescence values were obtained using a microplate reader. Lipid peroxidation measurement PAF-induced lipid peroxidation was measured using the BODIPY 581/591 C11 probe (S0043S; Beyotime). Briefly, cells plated in 12-well plates (5 × 104 cells/well) and 96-well black-walled plates (5 × 103 cells/well) were treated with PAF (60 µM) for 8 h. The cells were then incubated with 2 µM C11-BODIPY581/591 in Hank’s Balanced Salt Solution (1x HBSS, 14175–053 L; Gibco) for 30 min at 37 °C for lipid peroxidation measurements. After washing with PBS, fluorescence was visualized using a confocal microscope (CSIM110; SUNNY, China) and a microplate reader (Varioskan LUX Multimode Microplate Reader; Thermo Fisher Scientific). Lipid peroxidation was determined by calculating the ratio of oxidized (Ex/Em = 485/515 nm) to reduced (Ex/Em = 570/595 nm) C11 fluorescence signal. Transmission electron microscopy (TEM) The ultrastructure of mitochondria and autolysosomes was assessed using TEM. AN3CA and HEC1B cells were digested with 0.25% trypsin and fixed for 12 h in a 2.5% glutaraldehyde solution. After fixation, the samples were postfixed in 1% aqueous osmium tetroxide, dehydrated in a series of graded ethanol solutions (30% to 100%) with propylene oxide, embedded in epoxy resin, and incubated at 60 °C for 48 h. Semi-thin Sect. (1.0 mm) were stained with toluidine blue for visualization. Ultrathin sections of the cell samples were stained with 2% uranyl acetate and lead citrate. Ultrastructural images were captured using a TEM (HT7800; Hitachi, Tokyo, Japan). Immunofluorescence For cell immunofluorescence, AN3CA and HEC1B cells (5 × 105 cells/well) were seeded in confocal dishes and cultured overnight at 37 °C. The culture medium was then replaced with fresh medium containing PAF. After 48 h of incubation, the cells were washed, fixed, and blocked. Cells were then incubated with an antibody against ATF3 overnight at 4 °C. After incubation with Alexa Fluor 488 dye-conjugated secondary antibodies (P0176; Beyotime) at 25 °C for 1 h, the cells were incubated with Actin-Tracker Red-555 (C2203S; Beyotime) according to the manufacturer’s protocol. Subsequently, the cells were stained with DAPI and observed and photographed under a laser scanning confocal microscope (Nikon). Hematoxylin and eosin staining Paraffin-embedded xenograft tumor tissue samples were dewaxed and subjected to antigen retrieval. The sections were stained with Harris hematoxylin solution for 4 min, then differentiated using 0.8% hydrochloric acid alcohol for 2 s. Subsequently, they were stained with eosin solution (alcohol soluble) for 20 s, followed by dehydration and clearing. Images were acquired under a microscope (Olympus, Tokyo, Japan). IHC All experiments involving human tissues were approved by the Medical Ethics Committee of Guangzhou Women and Children’s Medical Center (Approval number: 124B00). Paraffin-embedded EC tumor tissue microarray (Cat. EC1601) or xenograft tumor tissue samples were dewaxed and subjected to antigen retrieval. The sections were blocked with 3% H2O2 at 25 °C for 10 min, followed by incubation with 10% goat serum for 30 min. Subsequently, the sections were incubated with the following primary antibodies overnight: PAF-AH2 (10085-1-AP, 1:20; Proteintech, China) for tumor microarrays; and Ki-67 (9129, 1:200; Cell Signaling Technology), glutathione peroxidase 4 (GPX4; ab125066, 1:2400; Abcam, Cambridge, UK), and 4-hydroxynonenal (4-HNE; ab46545, 1:300; Abcam) for xenograft tissues. The next day, the sections were incubated with a secondary antibody (ab205718, 1:2000; Abcam), followed by staining with a diaminobenzidine substrate chromogen mixture. Nuclear counterstaining was performed using hematoxylin for 1 min, and the slides were imaged using a microscope (Olympus). Bioinformatics analysis Ferroptosis-related genes were obtained from the FerrDb database (http://www.zhounan.org/ferrdb/current/) and Gene Ontology (https://www.geneontology.org/). We specifically focused on the genes categorized as “Drivers” and “Markers” within the FerrDb database, as well as those associated with the following Gene Ontology (GO) terms: cellular response to oxidative stress (GO:0034599), intracellular iron ion homeostasis (GO:0006879), and glutathione metabolic process (GO:0006749). All ferroptosis-related genes are listed in Additional file 2. After removing all duplicate genes, 270 ferroptosis-related genes were identified for subsequent analysis. Differentially expressed genes (DEGs) with |log2(fold change) | >0.5 and adjusted p-value < 0.05 were identified between the control and the PAF groups in HEC1B and AN3CA cells using the “limma” R package. The “ggplot2” R package was then employed to visualize the outcomes. Gene set enrichment analysis (GSEA) was subsequently performed using the “clusterProfiler” R package to investigate the association between the HEC1B and AN3CA cell lines and a customized ferroptosis-related pathway. Statistical analysis was conducted using R (v.4.4.3; R Foundation, Vienna, Austria) and the corresponding R package. Luciferase assay psiCHECK-2 vectors containing promoter sequences of the ATF3-binding site (pro-SLC7A11-WT/pro-GPX4-WT) and ATF3-binding site mutant (pro-SLC7A11-mut/pro-GPX4-mut) were constructed. For the reporter assays, cells seeded in 24-well plates (5 × 105cells/well) were transfected with either negative control (NC) siRNA or siRNA targeting ATF3 for 24 h. Cells were then subjected to transfection with the reporter plasmids and subsequent treatment with PAF. After 24 h, the cells were lysed, and reporter activity was measured using the Dual-Luciferase Reporter Assay System (E1910; Promega, Madison, WI, USA) following the manufacturer’s instructions. Molecular docking The AutoDock Vina 1.1.2 program was used to predict the binding mode of PAF (PubChem database, 108156) to ATF3 protein (AlphaFold database, P18847). The protein structure was prepared using AutoDock Vina, including the removal of water molecules and the addition of hydrogen atoms. The central coordinates (8.661, 8.512, -5.749) and a grid box size of 90 × 58 × 108 containing the entire protein were set. The PDBQT format was used to calculate the lowest binding affinity score, and the results were visually analyzed using PyMOL 2.5.2. Proteomic analysis For proteomic profiling, tissues were sent to Cosmos Wisdom Biotechnology (Hangzhou, China) for analysis. Fresh human normal endometrial tissue (CM) and human EC tissue samples (EC) were immediately frozen in liquid nitrogen overnight and stored at − 80 °C after harvest before testing. Informed consent for the use of leftover biological material for investigative purposes was obtained from all patients. Sequencing was performed according to the method described by Cosmos Wisdom, and the sequencing data were analyzed using R software. The proteomic analysis results are shown in Additional file 3. Statistical analysis Statistical analyses were performed using GraphPad Prism software (version 9.0; GraphPad Software Inc.). All experimental data are presented as the mean ± standard error of the mean from at least three independent experiments. For differences between groups, statistical significance was assessed via an unpaired two-tailed Student’s t-test or one-way ANOVA; two-way ANOVA was used for comparing multiple means across conditions; the Survival analysis was performed by applying the Kaplan-Meier method and the log-rank test. The chi-square test was used to conduct a correlation analysis between the expression level of PAF-AH2 and the TNM stage or tumor size of patients with EC. Statistical significance was set at p < 0.05. p > 0.05 is considered to be not significant and marked as ns. Results PAF inhibits EC cell proliferation in AN3CA and HEC1B cells in vitro To investigate the potential anti-cancer activity of PAF, we performed CCK8 experiments to assess the viability of AN3CA and HEC1B cells in response to treatment with PAF. We observed a dose-dependent reduction in the viability of EC cells 48 h after treatment with PAF exposure (Fig. 1A, B). Moreover, the colony formation assay revealed a decreased colony formation ability in the treated groups compared to that in the control group, indicating that the proliferation of AN3CA and HEC1B cells was significantly inhibited by PAF (Fig. 1C, D). Subsequently, we performed an EdU incorporation assay to assess the changes in DNA synthesis in PAF-treated AN3CA and HEC1B cells. Treatment with PAF led to a significant reduction in DNA synthesis compared to that observed in control cells (Fig. 1E, F). A PDO model can accurately recapitulate the molecular and phenotypic characteristics of human cancers. Therefore, to explore the potential clinical effects of PAF, we used tumor tissues from patients with EC to construct PDO models. Notably, following 48 h of treatment with PAF, we observed a fragmented organoid shape and significantly decreased ATP activity in the PDO group treated with a high PAF concentration (Fig. 1G, H). These results indicate that PAF effectively inhibited the proliferation of EC. PAF inhibits EC cell proliferation in AN3CA and HEC1B cells in vitro. (A, B) The IC50 values of PAF in AN3CA (A) and HEC1B (B) cells were determined using cell viability assays. One-way ANOVA with Tukey’s post hoc test was applied. (C, D) The colony-forming potential of AN3CA and HEC1B cells was assessed by crystal violet staining after incubation with 60 µM PAF. (E, F) PAF (60 µM) was incubated with AN3CA (E) and HEC1B (F) cells in a glass-bottom cell culture dish for 48 h. Proliferating cells were examined via EdU incorporation assay (orange) and Hoechst 33,342 (blue) nuclear staining for EC cells, and the images were merged. Scale bars, 100 μm. (G, H) Patient-derived EC organoids were treated with different concentrations of PAF (0, 20, 40, and 100 µM) for 48 h. Cell viability was measured using an ATP assay kit. Representative images (G) and statistical data (H) are shown. Scale bars, 200 μm. One-way ANOVA with Dunnett’s multiple comparisons test was used. (I) Representative transmission electron microscopy images show the mitochondrial morphology in AN3CA and HEC1B cells treated with control and PAF (60 µM) for 48 h. Scale bars: 2 μm (Black) or 1 μm (Green). ∗, p< 0.05; ∗∗∗, p <0.001 To further identify the morphological changes occurring in EC cells in response to treatment with PAF, we assessed cells using TEM. As depicted in Fig. 1I, TEM revealed that PAF-treated cells exhibited distinct mitochondrial morphological changes, including reduced mitochondrial size, a condensed mitochondrial matrix, and a decreased number of cristae. Collectively, these findings indicate that PAF effectively inhibited the proliferation of EC. PAF induces ferroptosis in EC cells To explore the mechanisms by which PAF inhibits EC proliferation, we performed RNA sequencing on control and PAF-treated EC cells. GSEA confirmed that the ferroptosis pathway was significantly activated in response to treatment with PAF (Fig. 2A, B). To provide a more intuitive visualization of the role of PAF in the regulation of ferroptosis, we compiled a customized gene set to include cellular responses to oxidative stress, intracellular iron ion homeostasis, glutathione metabolic processes, ferroptosis-promoting genes, and ferroptosis occurrence marker genes. The DEGs in the PAF-treated groups are presented in a volcano plot in Fig. 2C, D. GSEA revealed consistent enrichment patterns (Supplementary Fig. 1A, B), with PAF inducing a transcriptional state that closely resembled ferroptosis activation. Notably, there was a core (leading edge) enrichment of most signature genes involved in ferroptosis activation. PAF induces ferroptosis in EC cells. (A, B) Gene set enrichment analysis plots of the RNA-Seq analysis of ferroptosis-related gene modules in AN3CA (A) and HEC1B (B) cells treated with PAF. NES, normalized enrichment score. (C, D) Volcano plots of differentially expressed genes in the custom gene set. (E, F) Flow cytometry analysis of ROS levels in AN3CA (E) and HEC1B (F) cells treated with PAF for 48 h. (G) Intracellular Fe2+ intensity assays in AN3CA and HEC1B cells treated with PAF for 48 h. (H) BODIPY-C11 staining of AN3CA and HEC1B cells treated with PAF (60 µM) for 8 h. Scale bars, 100 μm. (I, J) Lipid peroxidation was measured using the BODIPY 581/591 C11 fluorescent probe. (K) WB analysis of the expression of SLC7A11, GPX4, and FTH1 in AN3CA and HEC1B cells treated with different concentrations (20, 40, and 60 µM) of PAF for 48 h. (L, M) AN3CA (L) and HEC1B (M) cell viability treated with a combination of PAF (60 µM) and the ferroptosis inhibitor ferrostatin-1 (Ferr-1, 10 µM). * p < 0.05, ** p < 0.01 and *** p < 0.001 We also investigated the effects of PAF on ferroptosis in EC cells. Intracellular ROS accumulation increased in a dose-dependent manner after treatment with PAF (Fig. 2E, F). Moreover, Fe2+ levels significantly increased after 48 h of treatment with PAF (Fig. 2G). Analysis of BODIPY 581/591 C11 probe fluorescence intensity confirmed that incubation with PAF increased lipid peroxidation (Fig. 2H–J). Moreover, we analyzed the expression patterns of ferroptosis-related markers in PAF-treated AN3CA and HEC1B cells. The expression of the anti-ferroptosis proteins FTH1, GPX4, and SLC7A11 was lower in the treated groups than in the control (Fig. 2K). We also exposed PAF-treated AN3CA and HEC1B cells to the ferroptosis inhibitor ferrostatin-1. This treatment significantly attenuated PAF-induced cell death in both cell lines, whereas treatment with inhibitors of apoptosis (Z-VAD-FMK) and necroptosis (Necrostatin-1, Necro) exerted no such effect (Fig. 2L, M; Supplementary Fig. 1C–F), supporting ferroptosis as the primary mechanism underlying the effect of PAF. PAF induces ferroptosis via the ATF3/GPX4 and ATF3/SLC7A11 axes As our results indicated that PAF inhibited cell proliferation by inducing ferroptosis, we further investigated the mediators through which ferroptosis was activated. Accumulating evidence has identified transcription factors (TFs) as critical regulators of ferroptosis [17]. Therefore, we then analyzed the transcriptome sequencing data in Fig. 2. The resulting heatmaps revealed differential expression of TFs between the control and the PAF-treated groups at the transcriptomic level in EC cell lines. Specifically, we identified DDIT3 and ATF3 as commonly differentially expressed TFs across EC cell lines (Fig. 3A). qRT-PCR further confirmed that treatment with PAF significantly upregulated the transcriptional levels of ATF3 and DDIT3 (Fig. 3B, C). However, WB analyses revealed treatment with PAF upregulated ATF3 protein expression (Fig. 3D, E) but not that of DDIT3 (Supplementary Fig. 2A). Moreover, immunofluorescence combined with confocal imaging of ATF3 protein in AN3CA and HEC1B cells demonstrated increased ATF3 expression in the PAF-treated groups (Fig. 3F). PAF induces ferroptosis via the ATF3/GPX4 and ATF3/SLC7A11 axes. (A) Heat map representing the significantly regulated transcription factors detected via RNA-seq analysis of AN3CA and HEC1B cells. (B, C) The mRNA expression levels of ATF3 and DDIT3 were in AN3CA and HEC1B cells treated with PAF (60 µM) for 48 h. (D, E) Protein expression of ATF3 in AN3CA and HEC1B cells treated with PAF (0, 20, 40, and 60 µM) for 48 h (D), and the quantifiable data are shown (E). (F) Confocal microscopy of AN3CA and HEC1B cells in the presence of PAF (60 µM). Cells were stained for ATF3 (green) and phalloidin (red). DAPI was used to visualize the nucleus (blue). Scale bar, 50 μm (left) or 100 μm (right). (G) The potential binding sites between ATF3 and GPX4/SLC7A11 were predicted by JASPAR. (H) Schematic diagram of ATF3 binding site-mutated SLC7A11 reporter vector (pro-SLC7A11-mut) and GPX4 reporter vector (pro-GPX4-mut). (I, J) Dual luciferase assay of the luciferase activity of AN3CA (I) and HEC1B (J) cells. (K) Immunoblotting analysis of the expression of GPX4 and SLC7A11 following ATF3 overexpression in AN3CA and HEC1B cells. (L) Immunoblotting analysis of the protein expression of SLC7A11, GPX4, and ATF3 after ATF3 silencing with siRNA coupled with PAF treatment in AN3CA cells. * p < 0.05, ** p < 0.01 and *** p < 0.001, ns. not significant Given that TFs regulate pathway activation by modulating the transcription of related genes, we analyzed publicly available ChIP-Seq data from the ENCODE Consortium. ATF3 binding to the promoter regions of SLC7A11 and GPX4 was observed in human HCT-116 (GEO: GSM1010757), A549 (GEO: GSM1010789), HepG2 (GEO: GSM803503), and K562 (GEO: GSM803380) cells (Supplementary Fig. 2B–E). Moreover, potential ATF3-binding sites approximately 2000 bp upstream of SLC7A11 and GPX4 were predicted using the JASPAR database (Fig. 3G). To further validate the binding relationship between ATF3 and the promoters of SLC7A11 and GPX4, we constructed luciferase reporter vectors containing wild-type promoters (pro-SLC7A11-WT and pro-GPX4-WT) and mutant versions (pro-SLC7A11-mut and pro-GPX4-mut) (Fig. 3H). Treatment with PAF significantly reduced luciferase activity in AN3CA and HEC1B cells. However, this effect was abolished both in the ATF3-knockdown groups and the ATF3-binding sites-mutating groups (Fig. 3I, J). In addition, ATF3 overexpression significantly decreased SLC7A11 and GPX4 protein levels in AN3CA and HEC1B cells (Fig. 3K). Notably, siRNA-mediated inhibition of ATF3 induced the expression of GPX4, whereas it did not significantly affect that of SLC7A11. Moreover, the inhibitory effects of PAF on SLC7A11 and GPX4 expression were reversed in the ATF3-knockdown groups (Fig. 3L). Furthermore, overexpression of SLC7A11 and GPX4 significantly reversed PAF-induced lipid peroxidation and cell death (Supplementary Fig. 2F–I). The findings collectively suggest that PAF promoted intracellular ATF3 accumulation, leading to the downregulation of SLC7A11 and GPX4 expression. PAF binds to ATF3 and decreases its ubiquitination Although PAF upregulated both the transcription and protein expression of ATF3, we focused on the post-translational regulation of ATF3, as its protein levels increased more rapidly than its RNA levels. To elucidate the mechanism underlying PAF-mediated upregulation of ATF3 expression, we performed molecular docking analysis to examine potential binding between PAF and ATF3. The three-dimensional and two-dimensional diagrams illustrate the binding between PAF and ATF3 (Fig. 4A, B). The analysis revealed a hydrogen bond formation between PAF and key amino acid residues (ARG-150 and GLU-166), along with hydrophobic interactions involving PHE-170 and LEU-169. A docking score provides insight into the binding affinity between a receptor and a ligand, with scores below − 4.25 kcal/mol indicating a possible binding and scores below − 6.9 kcal/mol indicating stronger binding [18]. For the PAF/ATF3 complex, the binding affinity score of − 7.0 kcal/mol indicated a strong molecular interaction. Furthermore, we conducted an SPR analysis to quantify the direct interaction between PAF and ATF3 using an anti-ATF3 antibody as a positive control. The results demonstrated robust binding interactions, evidenced by a KD of 3.24 µM (Fig. 4C, D). WB analysis of ATF3 revealed that the corresponding band almost completely disappeared in the control group at 50–65 °C, whereas it persisted in the PAF-treated cells (Fig. 4E). Furthermore, incubation with PAF enhanced ATF3 stability in the pronase-treated protein sample, suggesting PAF–ATF3 binding (Fig. 4F). We then treated AN3CA and HEC1B cells with cycloheximide (CHX), a protein synthesis inhibitor. ATF3 degradation was significantly reduced in PAF-treated groups compared to that in untreated controls, indicating that PAF binding stabilizes ATF3 (Fig. 4G). Moreover, treatment with MG132 (A proteasome inhibitor) reversed the CHX-induced decrease in protein levels, and the effect was enhanced by the co-administration of PAF (Fig. 4H). This result suggests that PAF regulates the stability of ATF3 through the ubiquitination pathway. Finally, the ubiquitination assay demonstrated that PAF significantly inhibited ATF3 ubiquitination in HEK293T cells (Fig. 4I). Collectively, these findings confirm that PAF enhanced ATF3 stability by preventing ubiquitination-mediated degradation. PAF binds to ATF3 and reduces its ubiquitination. (A, B) Molecular docking indicates the binding details between PAF and ATF3. The surface representation of the protein residues (A) and 2D representation of the binding interaction of PAF and ATF3 (B) are depicted. (C, D) Surface plasmon resonance of the affinity of anti-ATF3 antibody (C) and PAF (D) for ATF3 protein. KD, dissociation constant. (E) WB analysis shows that PAF stabilized ATF3 across different temperature gradients in the CETSA in 293T cells. (F) WB analysis indicates that PAF promoted the resistance of ATF3 to pronase digestion in the DARTS assay in 293T cells. (G) CHX chase analysis of ATF3 protein expression after treatment with PAF in AN3CA and HEC1B cells. (H) WB analysis of ATF3 in ATF3-overexpressing AN3CA and HEC1B cells. The cells were pretreated with PAF (60 µM) for 24 h and then treated with CHX and MG132 for 24 h. (I) Representative WB images demonstrate the ubiquitination of ATF3 in 293T cells co-transfected with ATF3-Flag, HA-Ub, and plasmids for 24 h. Cellular lysates were collected after 3 h of treatment with PAF, purified with a Flag-tag protein purification kit, and then subjected to WB with anti-HA and anti-ATF3 Treatment with PAF inhibits EC progression by promoting ferroptosis in vivo To further investigate the role of PAF in inducing ferroptosis in vivo, we constructed a xenograft model by subcutaneously injecting EC cells into BALB/c nude mice (Fig. 5A). Tumor proliferation was significantly attenuated in the PAF-treated group, as evidenced by imaging (Fig. 5B; Supplementary Fig. 3A, B), tumor volume measurements (Fig. 5C), and tumor weight analysis (Fig. 5D), while body weight exhibited no significant differences (Supplementary Fig. 3C, D). HE staining revealed that tissue sections from the PAF-treated group exhibited scattered cellular distribution with reduced density, cellular rupture, and prominent nuclear pyknosis, demonstrating the anti-cancer activity of PAF (Fig. 5E). Consistent with these findings, immunohistochemical staining for Ki67, a well-established proliferation marker, revealed reduced expression in PAF-treated tumors compared to vehicle controls (Fig. 5F, G). To further explore the relevance of PAF-induced tumor growth inhibition and ferroptosis, we performed immunohistochemical analysis to detect the expression of critical targets of ferroptosis, namely 4-HNE and GPX4, in tumor tissues. Notably, treatment with PAF considerably increased the levels of 4-HNE, a stable lipid peroxidation byproduct and established ferroptosis biomarker. Furthermore, GPX4 protein expression was significantly downregulated in PAF-treated tumors (Fig. 5H, I). Collectively, these data indicate that treatment with PAF suppressed EC progression by inducing ferroptosis in vivo. Treatment with PAF inhibits EC progression by promoting ferroptosis in vivo. (A) Timeline of the nude mice experiments. Mice bearing AN3CA and HEC1B subcutaneous tumors were injected with PAF every 3 days for a total of 6 times. (B) The resected tumors. (C, D) Volume (C) and weight (D) of tumors. The negative control group received a saline solution (vehicle). Data are presented as means ± SEM and were analyzed using the Student’s t-test was applied (n = 5 in the AN3CA group; n = 6 in the HEC1B group). (E) Hematoxylin and eosin staining of tumors from mice in different groups. Scale bar: 200 μm. (F) Representative images of immunohistochemistry analysis of Ki67 expression in mouse tumor samples. Scale bars, 200 μm. (G) Percentage of Ki67-positive stained cells per field. (H) Immunohistochemical staining for GPX4 and 4-HNE in tissue collected from AN3CA and HEC1B xenografted tumors. Scale bars: 200 μm (Black) or 100 μm (Red). (I) Average optical density of proteins (assessed with ImageJ software). Statistical graphs of immunohistochemistry staining for GPX4 and 4-HNE in tumor xenografts from the indicated groups. * p < 0.05, ** p < 0.01 and *** p < 0.001 Endometrial cancerization is accompanied by impaired PAF metabolism PAF metabolism is tightly regulated by a series of enzymes, such as PAF-AHs, which can enzymatically remove acetyl groups at the sn-2 position, resulting in the production of lyso-phosphatidylcholine, the biologically inactive form of PAF. In contrast, lyso-phosphatidylcholine acyltransferases (LPCAT1 and LPCAT2) are key biosynthetic enzymes in the PAF remodeling pathway [19, 20]. Therefore, we investigated proteome abundance in normal human endometrial and EC tissue samples to assess the internal PAF metabolism. We detected a total of 8128 proteins, of which 904 (524 downregulated and 380 upregulated) were significantly differentially regulated in EC tissue samples (Fig. 6A). Analysis of the differentially regulated proteins in EC tissue samples compared to those in normal endometrial tissues revealed that many of these proteins were involved in catalytic and enzyme regulatory activities (Supplementary Fig. 4A), indicating heightened metabolic activity in the tumor microenvironment. Therefore, we focused on the hydrolase and transferase activities in GO “child terms” of catalytic activity. GO molecular function analyses revealed a significant enrichment in carboxylic ester hydrolase activity, acetyltransferase activity, and acyltransferase activity (Supplementary Fig. 4B, C). Notably, heatmap visualization of proteins associated with these functions identified upregulated PAF-AH2 expression in EC tissues (Fig. 6B, C). However, the levels of LPCAT1 and LPCAT2 did not differ significantly compared to those in the normal human endometrial tissue (Fig. 6D). Subsequent analysis of EC tissue microarrays revealed a significantly elevated PAF-AH2 staining intensity in tumor tissues compared to adjacent normal tissues (Fig. 6E, F). Moreover, elevated PAF-AH2 protein levels were correlated with poor survival in patients with EC (Fig. 6G). Notably, high PAF-AH2 expression was prominently detected in EC tissues at stage III/IV (Fig. 6H). We also observed that high PAF-AH2 expression was strongly correlated with the tumor size in EC tissues (Fig. 6I). Furthermore, PAF-AH2 overexpression significantly reversed PAF-induced lipid peroxidation and cell death (Supplementary Fig. 4D–H). These results suggest that suppressing the upregulated PAF-AH2 expression may confer great benefits to patients with EC. EC is accompanied by abnormal PAF metabolism. (A) Volcano plots of the identified proteins. Each dot represents one protein. Significantly altered proteins were identified based on a fold change cut-off (1.5 or -1.5-fold), and a p-value of < 0.05, determined using the t-test. (B, C) Heatmap showing the differentially expressed proteins in transferase activity (B) and hydrolase activity (C). (D) Expression of PAF-AH2, LPCAT1, and LPCAT2 in different groups. (E, F) Representative micrographs of immunohistochemical staining for PAF-AH2 in EC and adjacent normal tissues (E). Differences in immunohistochemistry scores between EC and adjacent normal tissues (n = 80 pairs). The differences were determined using non-parametric tests (F). Scale bars, 400 μm (left), 50 μm (right). (G) Kaplan–Meier survival analysis and log-rank statistical tests of the correlation between PAF-AH2 expression levels and overall survival in patients with EC. (H) Quantitation of PAF–AH2 expression in 80 EC tissues with different TNM stages according to the IHC scores. (I) Quantitation of PAF-AH2 expression in EC tissues with different tumor sizes according to the IHC score. CM: Human normal endometrial tissue, EC: Human EC tissue samples. ns. not significant Discussion PAF, also known as 1-alkyl-2-acetyl-sn-glycero-3-phosphocholine, is a highly potent phospholipid-derived biological mediator involved in several important biological functions, including platelet aggregation and allergic reactions [20]. The biological functions are mediated by platelet-activating factor receptors (PAFRs), which play pivotal roles in tumor progression, including oncogenic transformation, anti-apoptosis, and metastasis [21]. Notably, PAF can exhibit either tumor-promoting or suppressive effects depending on the cellular or tissue context [22, 23]. In the present study, we demonstrated that PAF exerts notable inhibitory effects on the proliferation of the EC cell lines AN3CA and HEC1B in vitro. We further confirmed the anti-tumor effect of PAF in EC xenograft models and PDO models. We also elucidated the intricate regulatory relationship between ATF3 and PAF in ferroptosis, revealing an important regulatory module that controls ferroptosishomeostasis in endometrial tissues (Fig. 7). Finally, we demonstrated that PAF induces ferroptosis and inhibits EC progression both in vitro and in vivo by binding to ATF3 and subsequently suppressing the transcription of ferroptosis-related genes. Collectively, our results reveal the potential role and mechanism of PAF in EC. Schematic overview of the present study. Treatment with PAF attenuates EC cell proliferation by inducing ferroptosis. Mechanistically, PAF interacts with ATF3, enhancing its stability through deubiquitination and consequently suppressing the expression of the ferroptosis-related proteins SLC7A11 and GPX4 Ferroptosis, an iron-dependent programmed cell death characterized by lipid peroxidation and oxidative stress, has emerged as a promising therapeutic target for various malignancies [24, 25]. Accumulating evidence suggests a strong link between ferroptosis and PUFA lipids. P450 oxidoreductase promotes ferroptosis by upregulating the peroxidation of membrane polyunsaturated phospholipids [26]. Moreover, research on ferroptosis susceptibility has revealed that the synthesis of polyunsaturated ether phospholipids contributes to ferroptosis, with the pro-ferroptotic effect specifically mediated through ether lipid biosynthesis [27]. Collectively, these findings highlight the relationship between lipid metabolism and ferroptosis. In this study, our transcriptome sequencing data indicated that PAF induces ferroptosis in EC cell lines. Treatment with PAF increased the levels of ROS, free iron, and lipid peroxidation. In addition, PAF increased the levels of 4-HNE and suppressed the expression of GPX4 in vivo. Overall, these findings suggest that PAF induces ferroptosis in EC. Recent studies have highlighted the role of TFs in controlling the activity of oncogene transcription, which influences different aspects of tumor growth [28, 29]. For example, hypoxia-inducible factor-1α is a key activator of the multidrug resistance gene 1, which encodes a membrane transporter protein responsible for cellular drug efflux [30]. In the present study, we focused on the expression of TFs using RNA-seq data and found that the transcription of ATF3 and DDIT3 was upregulated in both AN3CA and HEC1B cells in response to treatment with PAF. Moreover, WB and cell immunofluorescence analyses revealed increased ATF3 expression in the PAF-treated group. We subsequently validated this hypothesis using molecular docking, SPR, CETSA, and DARTS. Furthermore, subsequent ubiquitination analysis confirmed that PAF inhibited ATF3 ubiquitination. ATF3 belongs to the ATF/cAMP response element binding family and is a basic leucine zipper protein. It plays a critical role in regulating various biological processes, including endoplasmic reticulum stress, oxidative stress, and metabolic regulation [31, 32]. As an adaptive-response gene, the regulation of ATF3 enables cells to adapt to external or internal changes by activating or inhibiting gene expression via signaling from various receptors. ATF3 promotes ferroptosis by binding to the SLC7A11 promoter and suppressing system Xc−, leading to reduced intracellular glutathione levels and increased cellular sensitivity to RSL3-induced GPX4 inhibition [33]. Our analysis of publicly available ChIP-Seq data from the ENCODE Consortium revealed ATF3 binding peaks at the promoter regions of SLC7A11 and GPX4. Consistently, luciferase reporter assays revealed that treatment with PAF downregulated the transcriptional expression of SLC7A11 and GPX4 in an ATF3-dependent manner. Moreover, ATF3 overexpression significantly decreased SLC7A11 and GPX4 protein expression levels, whereas its downregulation did not induce SLC7A11 expression as expected. Notably, SLC7A11 is also regulated by multiple transcription factors, such as ATF4 and nuclear factor erythroid 2-related factor 2 (NRF2) [34, 35]. Given that basal ATF3 expression is inherently low in AN3CA cells, the downregulation of ATF3 is expected to exert a negligible effect on SLC7A11 expression. Nevertheless, these findings suggest ATF3 as an important regulator of PAF-induced ferroptosis. Given the low concentration and short biological half-life of PAF, accurate quantification of its concentrations remains challenging [8]. PAF metabolism is mediated by specific enzymes. In this study, we found that elevated PAF-AH2 expression in patients with EC was significantly associated with poorer survival. A previous study has demonstrated that blocking PAF-AH2 remodels the lipidome toward a ferroptosis state characterized by the accumulation of ether-linked phosphatidylethanolamines and other pro-ferroptotic phospholipids [36]. Our findings suggested that enhanced PAF degradation in EC may afford tumors an escape route from ferroptosis. Notably, this microenvironment, which facilitates the rapid catabolism of PAF potentially accounts for prior observations of elevated lyso-PAF concentrations in the peripheral blood of patients with EC [6]. This study has certain limitations that should be acknowledged. First, the nature of the sample analyzed is limited. To confirm the clinical utility of our findings, future studies using a larger cohort of clinical samples for PDO models and proteomic analysis are warranted. Second, although we confirmed that PAF-AH2 protein expression is upregulated in EC tissues, its contribution to EC progression and the underlying mechanisms remain unclear. In conclusion, we demonstrated that PAF inhibits EC cell proliferation by inducing ferroptosis. Mechanistically, PAF interacts with ATF3, enhancing its stability through deubiquitination and consequently suppressing the transcription of the ferroptosis-related genes SLC7A11 and GPX4. In addition, EC tissues exhibit upregulated PAF-AH2 expression. Overall, this study highlights PAF metabolism as a promising direction for EC diagnostics and therapeutics. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: The RNA sequencing result of AN3CA and HEC1B. Supplementary Material 2: Customized gene sets (Ferroptosis-related genes). Supplementary Material 3: The proteomic analysis results of human tissues. Supplementary Material 4: The specific information of the patients involved in this study. Supplementary Material 5: The full uncropped Western Blot images. Supplementary Material 6: Fig. 1. PAF induces ferroptosis in EC cells. (A, B) Gene set enrichment analysis of cell ferroptosis activation signature in AN3CA (A) and HEC1B (B) cells after treatment with PAF (Versus control). (C–D) The percentage of AN3CA (C) and HEC1B (D) cell viability after treatment with a combination of PAF (60 μM) and the necrosis inhibitor necrostatin-1 (Necro-1, 10 μM) for 48 h. (E–F) AN3CA (E) and HEC1B (F) cell viability after treatment with a combination of PAF (60 μM) and the apoptosis inhibitor Z-VAD-FMK (10 μM) for 48 h. ns. not significant. Supplementary Material 7: Fig. 2. PAF induces ferroptosis via the ATF3/GPX4 and ATF3/SLC7A11 axes (A) Protein expression of DDIT3 in AN3CA and HEC1B cells treated with PAF (0, 20, 40, and 60 µM) for 48 h. (B–E) Visualization of ATF3 in CHIP-seq profile within the genomic region surrounding the SLC7A11 (B), ACSL4 (C), GPX4 (D), and FTH1 (E) loci. (F) Representative western blot images of SLC7A11 in HEC1B cells overexpressing SLC7A11. (G) Representative western blot images of GPX4 in HEC1B cells overexpressing GPX4. (H) Lipid peroxidation was measured using the BODIPY 581/591 C11 fluorescent probe in HEC1B cells in response to PAF (60 µM) for 24 h. (I) The effect of PAF (60 µM) on the viability of HEC1B cells after being treated for 24 h. Supplementary Material 8: Fig. 3. Treatment with PAF inhibits EC progression in vivoin vivo. (A, B) Representative images of AN3CA (A) and HEC1B (B) xenograft tumors after PAF treatment for 15 days. (C, D) The change in body weight of mice was presented. ns. not significant. Supplementary Material 9: Fig. 4. Overexpression of PAF-AH2 attenuates PAF-Induced ferroptosis (A) Graphical representation of enrichment analysis results based on the significantly altered proteins. Most enriched Gene Ontology molecular function terms are enriched. (B, C) GO molecular function analyses of the significantly altered proteins. The top enrichment pathways, transferase activity (B) and hydrolase activity (C), are presented. (D) Representative western blot images of PAF-AH2 in AN3CA and HEC1B cells overexpressing PAF-AH2. (E, F) Lipid peroxidation was measured using the BODIPY 581/591 C11 fluorescent probe in AN3CA (E) and HEC1B (F) cells in response to PAF (60 µM) for 24 h. (G, H) The effect of PAF (60 µM) on the viability of AN3CA (G) and HEC1B (H) cells after being treated for 24 h. Acknowledgements Not applicable. Abbreviations Endometrial carcinoma Platelet-activating factor Activating transcription factor 3 Immunohistochemical Hematoxylin and eosin PAF-acetylhydrolase 2 Surface plasmon resonance Phosphocholine Lyso-platelet-activating factor-16 5-ethynyl-20-deoxyuridine Western blot Patient-derived organoid Reactive oxygen species Cellular thermal shift assay Drug affinity responsive target stability Transmission electron microscopy Gene Ontology Differentially expressed genes Gene set enrichment analysis Negative control Adenosine triphosphate Dissociation constant Cycloheximide PAF receptors Author contributions Feifei Pan conceived and designed the study, conducted experiments, analyzed data, and drafted the manuscript. Yuqing Qiu, Yanqin Zheng, Yun Chen, Xiaojie Liu, and Zixin Tao assisted with the experimental work and data collection. Zhuojie Huang and Jingyao Wang performed the bioinformatics analysis. Yiming Song, Lirong Guo, Yaqiong Liu, and Xingcui Xuan collected the clinical samples. Zhe-Sheng Chen, Kunxiang Gong and Kun Shi provided guidance on experimental design and data interpretation. All authors read and approved the final manuscript. Funding This work was supported by the National Natural Science Foundation of China (grant numbers 82072859 to K.S. and 82404926 to K.X.G.), and Guangdong Basic and Applied Basic Research Foundation (grant numbers 2022A1515220127 to K.S. and 2025A1515011310 to K.X.G.). Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval and consent to participate Ethics approval and consent to participate in the acquisition and use of human specimens were in concordance with the International Ethical Guidelines for Research Involving Human Subjects, as stated in the Helsinki Declaration, and were approved by the Medical Ethics Committee of Guangzhou Women and Children’s Medical Center (Approval number: 124B00). The specific information of the patients involved in this study was presented in Additional file 4. The animal experiments were approved by the Guangzhou Women and Children’s Medical Center Animal Welfare and Ethics Committee (Approval number: RSDW­2023­01427) and were conducted in accordance with the NIH Guidelines for the Care and Use of Laboratory Animals. Consent for publication Not applicable. Patient consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Kunxiang Gong, Email: gkx13265040169@126.com. Kun Shi, Email: shikun28@hotmail.com. References Associated Data Supplementary Materials Supplementary Material 1: The RNA sequencing result of AN3CA and HEC1B. Supplementary Material 2: Customized gene sets (Ferroptosis-related genes). Supplementary Material 3: The proteomic analysis results of human tissues. Supplementary Material 4: The specific information of the patients involved in this study. Supplementary Material 5: The full uncropped Western Blot images. Supplementary Material 6: Fig. 1. PAF induces ferroptosis in EC cells. (A, B) Gene set enrichment analysis of cell ferroptosis activation signature in AN3CA (A) and HEC1B (B) cells after treatment with PAF (Versus control). (C–D) The percentage of AN3CA (C) and HEC1B (D) cell viability after treatment with a combination of PAF (60 μM) and the necrosis inhibitor necrostatin-1 (Necro-1, 10 μM) for 48 h. (E–F) AN3CA (E) and HEC1B (F) cell viability after treatment with a combination of PAF (60 μM) and the apoptosis inhibitor Z-VAD-FMK (10 μM) for 48 h. ns. not significant. Supplementary Material 7: Fig. 2. PAF induces ferroptosis via the ATF3/GPX4 and ATF3/SLC7A11 axes (A) Protein expression of DDIT3 in AN3CA and HEC1B cells treated with PAF (0, 20, 40, and 60 µM) for 48 h. (B–E) Visualization of ATF3 in CHIP-seq profile within the genomic region surrounding the SLC7A11 (B), ACSL4 (C), GPX4 (D), and FTH1 (E) loci. (F) Representative western blot images of SLC7A11 in HEC1B cells overexpressing SLC7A11. (G) Representative western blot images of GPX4 in HEC1B cells overexpressing GPX4. (H) Lipid peroxidation was measured using the BODIPY 581/591 C11 fluorescent probe in HEC1B cells in response to PAF (60 µM) for 24 h. (I) The effect of PAF (60 µM) on the viability of HEC1B cells after being treated for 24 h. Supplementary Material 8: Fig. 3. Treatment with PAF inhibits EC progression in vivoin vivo. (A, B) Representative images of AN3CA (A) and HEC1B (B) xenograft tumors after PAF treatment for 15 days. (C, D) The change in body weight of mice was presented. ns. not significant. Supplementary Material 9: Fig. 4. Overexpression of PAF-AH2 attenuates PAF-Induced ferroptosis (A) Graphical representation of enrichment analysis results based on the significantly altered proteins. Most enriched Gene Ontology molecular function terms are enriched. (B, C) GO molecular function analyses of the significantly altered proteins. The top enrichment pathways, transferase activity (B) and hydrolase activity (C), are presented. (D) Representative western blot images of PAF-AH2 in AN3CA and HEC1B cells overexpressing PAF-AH2. (E, F) Lipid peroxidation was measured using the BODIPY 581/591 C11 fluorescent probe in AN3CA (E) and HEC1B (F) cells in response to PAF (60 µM) for 24 h. (G, H) The effect of PAF (60 µM) on the viability of AN3CA (G) and HEC1B (H) cells after being treated for 24 h. Data Availability Statement No datasets were generated or analysed during the current study.