MAP4K4 inhibition enhances the anti-ferroptosis effect of the FTO/YAP/ATF4 axis in doxorubicin-induced cardiotoxicity

Abstract

Background

Doxorubicin (DOX) is a highly effective chemotherapeutic agent, but its clinical application was constrained by irreversible cardiotoxicity. Increasing evidence indicates that ferroptosis is a central mechanism of DOX-induced cardiotoxicity (DIC). Both the m⁶A demethylase FTO and yes-associated protein (YAP) are found to suppress ferroptosis in DIC. Importantly, both YAP and its upstream negative regulator, MAP4K4, are potential targets of FTO. However, whether FTO regulates DOX-induced cardiac ferroptosis via the MAP4K4/YAP signaling axis remains unexplored.

Methods

FTO and YAP expressions were analyzed in DOX-treated cardiac tissues and cardiomyocytes. The effect of FTO on ferroptosis, as well as on the expression of YAP, ATF4, and MAP4K4, was evaluated in DOX-treated mouse heart and cardiomyocytes using cardiac-specific FTO AAV delivery or lentiviral transfection. Pharmacological inhibition of YAP, ATF4, or MAP4K4 was further performed to investigate the underlying mechanisms. In addition, the m⁶A modifications of FTO on MAP4K4 and YAP transcripts were examined.

Results

FTO and YAP were significantly reduced in DOX-treated cardiac tissues and cardiomyocytes. FTO overexpression markedly attenuated DIC by inhibiting ferroptosis by enhancing YAP and ATF4 expressions. Inhibition of YAP or ATF4 significantly abolished the protective and anti-ferroptotic effects of FTO. FTO also increased MAP4K4 expression, whereas MAP4K4 inhibition further enhanced the anti-ferroptotic effects mediated by the FTO/YAP/ATF4 axis. Mechanistically, FTO promoted m⁶A demethylation of MAP4K4 and YAP transcripts and enhanced their mRNA stability.

Conclusions

FTO protects against DIC by inhibiting ferroptosis via up-regulation of the YAP/ATF4 axis, and inhibition of MAP4K4 further enhances the FTO/YAP/ATF4-mediated cardioprotective and anti-ferroptotic effects.

Graphical Abstract

• DOX increases m⁶A methylation and suppresses FTO/YAP expression in the heart.

• FTO attenuates DOX-induced ferroptosis via m⁶A-dependent upregulation of YAP.

• FTO/YAP activates ATF4 signaling to confer cardioprotection against ferroptosis.

• FTO upregulates MAP4K4, while MAP4K4 inhibition amplifies FTO/YAP/ATF4-mediated protection.

Introduction

Doxorubicin (DOX) is a highly effective chemotherapeutic agent; however, its cumulative dosage cardiotoxicity can ultimately lead to irreversible myocardial injury and heart failure, thereby limiting its clinical application (Berecz et al. 2022). To date, dexrazoxane, an iron-chelating agent, remains the only clinical agent used to mitigate DOX-induced cardiotoxicity (DIC) (Byun et al. 2019). Nevertheless, its clinical use is restricted to patients receiving a cumulative DOX dose exceeding 300 mg/m² (Chen et al. 2020). Ferroptosis, a non-apoptotic regulated cell death driven by iron-dependent lipid peroxidation, has emerged as a key contributor to various pathological conditions (Chen et al. 2025; Chow et al. 2023). Growing evidence identifies ferroptosis as a central mechanism of DIC, as the ferroptosis inhibitor ferrostatin-1 exhibits stronger cardioprotective effects than inhibitors targeting autophagy, apoptosis, or necroptosis (de Baat et al. 2022). Therefore, targeting ferroptosis has become a promising therapeutic approach for preventing DIC.

Recent studies have increasingly recognized that Yes-associated protein (YAP) is a crucial regulator of cardiac ferroptosis and injury (Fang et al. 2019; Fiedler et al. 2019). RNA sequencing combined with experimental validation revealed that YAP expression is markedly decreased in failing human hearts and in models of DIC (Gao et al. 2021; Golforoush et al. 2020). Pharmacological activation of YAP by isorhapontigenin or melatonin markedly attenuates DOX-induced ferroptosis and cardiotoxicity (Gao et al. 2021; Golforoush et al. 2020). Increasing evidence indicates that YAP suppresses ferroptosis by maintaining mitochondrial function, redox homeostasis, and iron metabolism (Govender et al. 2014; Han et al. 2021; He et al. 2023). Mechanistically, YAP promotes the transcription of antioxidant-related genes, including nuclear factor erythroid 2-related factor 2 (Nrf2) and activating transcription factor 4 (ATF4) by cooperating with TEAD transcription factors (Heallen and Martin 2023; Identification 2001). In line with this, the Nrf2/ATF4 signaling axis activated by the methyltransferase inhibitor BRD4770 plays a critical role in alleviating DIC through induction of the SLC7A11 (xCT) pathway (Jiang et al. 2022). Moreover, irisin-mediated activation of ATF4 significantly protects against DIC (Jin et al. 2025). This evidence demonstrates the importance of the YAP/ATF4 axis in cardiac ferroptosis in DIC.

N6-methyladenosine (m⁶A) modification has also emerged as an important regulator of DOX-induced cardiac ferroptosis (Ke et al. 2022). Previous studies showed that DOX treatment upregulated the m⁶A methyltransferases METTL3 and METTL14, thereby promoting cardiac ferroptosis (Kitakata et al. 2022). In contrast, our previous work demonstrated that the m⁶A demethylase FTO protected against DIC by inhibiting ferroptosis (Li et al. 2025). Moreover, m⁶A modification is also involved in regulating YAP expression in cardiovascular diseases. For example, FTO alleviates ischemia–reperfusion (I/R)-induced cardiac ferroptosis by mediating m⁶A demethylation of YAP transcripts (Matsuda et al. 2016). Additionally, METTL3 mediates m⁶A modification of large tumor suppressor kinase (LATS), which promotes YAP phosphorylation and degradation. MAP4K4, another upstream regulator of YAP phosphorylation, has also been implicated in DIC and YAP signaling (Meng et al. 2015, 2023). Importantly, pharmacological inhibition of MAP4K4 improves cardiomyocyte function and reduces myocardial infarction sizes (Ndubaku et al. 2015). Furthermore, FTO has been reported to up-regulate MAP4K4 expression in gastric cancer through m⁶A modification (Qiu et al. 2023).

In this study, we assayed whether FTO regulates ferroptosis by modulating MAP4K4 and YAP/ATF4, aiming to elucidate the underlying mechanism and identify a more effective therapeutic strategy to mitigate DIC.

Methods

Animal experiment

Male C57BL/6 J mice (6–8 weeks) were sourced from Weitong Lihua (Beijing, China). Mice were intraperitoneally injected with a single dosage of 25 mg/kg DOX (MCE, New Jersey, USA) to establish an acute DIC model (Qiu et al. 2023), and received tail vein injections of AAV9-cTnT-FTO (1 × 10¹¹ vector genomes per mouse) for 4 weeks for cardiac-specific FTO overexpression before DOX administration. Animal manipulations strictly complied with the principles of the Guide for the Care and Use of Laboratory Animals.

Echocardiography

Mice were anesthetized using 1.25% (w/v) tribromoethanol, followed by removal of chest hair. Transthoracic echocardiography was conducted using a MYLAB™ X8 VET ultrasound system (Suzhou, China). Cardiac systolic function was evaluated by determining left ventricular ejection fraction (LV-EF) and fractional shortening (LV-FS).

Heart tissue collection

Mice were subjected to deep anesthesia via CO₂ inhalation at a continuous flow rate of 1.5 L/min, followed by rapid cardiac resection. Excised hearts were perfused with pre-cooled phosphate-buffered saline (PBS) to flush out residual blood inside the cardiac vasculature. Cardiac tissues were subsequently collected and processed for transmission electron microscopy, histological staining, and western blot analysis.

Histopathology

Harvested cardiac specimens underwent 24-h fixation in 10% neutral buffered formalin solution. Afterwards, these samples were embedded in paraffin wax and prepared into slices (4 nm). Hematoxylin and eosin (H&E) staining was then conducted to visualize pathological morphological changes of the myocardium.

Transmission Electron Microscopy (TEM)

To evaluate myocardial ultrastructural alterations under DOX treatment, heart tissues were fixed using a mixed fixative containing 2.5% glutaraldehyde and 1% osmium tetroxide. Fixed specimens were processed through gradient ethanol dehydration, resin embedding, and ultrathin slicing (70 nm). Upon sequential double staining with uranyl acetate and lead citrate, specimen slices were imaged and documented under a TEM.

Cell culture and transfection

H9c2 cells were cultured with DMEM containing 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin in a humidified incubator with 5% CO₂ at 37 °C. H9c2 cells were transfected with lentiviral vectors encoding FTO (oe FTO), specific shRNA (sh-FTO), or negative control vector (NC) (GeneChem Co., Ltd.,Shanghai) and then treated with puromycin (5 μg/ml) for 7 days. To assay the role of YAP/ATF4 in the anti-ferroptosis effect of FTO, cells were also transfected with YAP or ATF4-specific small interfering RNA (si-YAP, sense:5’-UGAAACAGCAGGAGUUAAUUTT-3’; antisense:5’-AAUAACUCCUGCUGUUUCATT-3’); (si-ATF4, sense:5’-GCUGCUUAUAUACUCUAATT-3’; antisense:5’-UUAGAGUAUAUAAGCAGCTT-3’), or si-negative control (si-NC) (sense:5’-UUCUCCGAACGUGUCACGUdTdT-3’; antisense:5’-ACGUGACACGUUCGGAGAAdTdT-3’).

All siRNA duplexes were transfected via Lipofectamine™ 2000 (Invitrogen, Carlsbad, CA, USA) and sustained for 48 h after transfection for downstream functional detection.

Cell Counting Kit-8 (CCK-8) assay

Cell viability was evaluated to determine the effects of FTO/YAP on DIC. Briefly, H9c2 cells were seeded into 96-well plates at a density of 5 × 10³ cells per well and exposed to DOX for 24 h. Subsequently, cells were incubated with CCK-8 working solution for 2 h, and absorbance at 450 nm was measured using a microplate reader.

Intracellular Fe²⁺ measurement

To determine Fe²⁺ levels in H9c2 cells following DOX treatment, 5 μM FerroOrange fluorescent probe (Dojindo, Japan), a highly selective fluorescent indicator, was added to the cell culture and stained for 30 min. Afterwards, the excess unbound probe was thoroughly washed using PBS. Cellular fluorescent intensity was subsequently detected with Olympus imaging equipment at an excitation (Ex) wavelength of 488 nm and an emission (Em) wavelength of 530 nm (Olympus Corporation).

Intracellular Reactive Oxygen Species (ROS) analysis

To assay the effect of DOX on ROS levels in H9c2 cells, cells were incubated with 2.5 μM dihydroethidium (DCFH; Beyotime Biotechnology) for 30 min. After washing with PBS to remove excess probe, intracellular superoxide production was evaluated by fluorescence at Ex/Em wavelengths of 488/530 nm.

Mitochondrial membrane potential (MMP) assay

To assay the effect of DOX on mitochondrial function, H9c2 cells were incubated with the JC-1 staining solution for 20 min, and then washed to remove residual dye. Dual-channel fluorescence signals were captured at two sets of optical parameters: red fluorescence (Ex/Em:525/590 nm) and green fluorescence (Ex/Em:490/530 nm). Mechanistically, JC-1 accumulates within polarized mitochondria as aggregates that emit red fluorescence, whereas loss of mitochondrial membrane potential (MMP) prevents aggregation, resulting in monomer formation and green fluorescence emission.

Detection of YAP cellular distribution by immunofluorescence

To confirm the regulation of MAKP4K4 on YAP in H9c2 cells following DOX treatment, cells were exposed to DMX-5804 or DOX for 24 h prior to fixation with pre-cooled acetone (10 min). Cells were permeabilized in 0.3% Triton X-100 and blocked with 5% BSA. The specimens were then sequentially incubated with a YAP-specific primary antibody and corresponding fluorescent secondary antibodies. After washing to remove the antibodies, cells were further stained with DAPI; Fluorescence intensity was detected at 488/530 nm (green) and at 405/450 nm (blue).

Quantitative Reverse Transcription PCR (qRT-PCR) assay

Cellular total RNA was isolated and reverse-transcribed into single-stranded cDNA using the TransGen kit. FTO was analyzed by qRT-PCR tests using a thermal cycling protocol that included an initial denaturation step, followed by amplification and quantification cycles. Fluorescence was measured at each amplification cycle, and threshold cycle (Ct) values were obtained for the target gene assay (TransGen Biotech). The FTO-specific primers were listed below: Forward primer: 5’-GCTGTGGAAGAAGATGGAGAGTGTG-3’; Reverse primer: 5’-AGTGGGGATCAGGACGGGCAGAC-3’.

m⁶A quantification assay

m⁶A RNA Methylation Quantification Kit (EpiGentek) was used for assaying m⁶A RNA level. Briefly, total RNA was first extracted and purified as described above. Each reaction well was loaded with 200 ng RNA specimen plus 2 μl positive/negative control solutions, and optical absorbance was recorded at 450 nm using a microplate spectrophotometer. The proportional level of m⁶A was computed via the listed formula:

$${\text{m}}^6\text{A}\left(\text{\%}\right)=[(\text{PC}-\text{NC})/(\text{Sample}-\text{NC})]\times (\text{P}/\text{S})\times 100\text{\%}$$

PC OD, NC OD, and Sample OD correspond to optical density readings of positive control, negative control, and test sample separately; P and S refer to the initial loading quantity of positive control RNA and sample RNA individually.

M6A RNA immunoprecipitation assay (MeRIP)

MeRIP-qPCR was performed to determine the m⁶A methylation modification of FTO on MAP4K4 or YAP mRNA. In brief, total RNA was extracted from DOX-treated H9c2 cells overexpressing FTO or the negative control, followed by mRNA purification using kits from BersinBio. After purification, the mRNA was chemically fragmented into short transcripts averaging approximately 300 nucleotides in length by treatment with ZnCl₂ in an ice-water mixture for 10 s. 10% of the RNA sample was retained as input, whereas the remaining RNA was used for immunoprecipitation with the Magna ChIP kit. Protein A/G Magnetic Beads pre-bound to either an anti-m⁶A antibody or a normal mouse IgG antibody. Following immunoprecipitation, MAP4K4 or YAP RNA was analyzed by qPCR, with results normalized to the input RNA. Gene-specific primers were designed based on the putative m⁶A motifs predicted by SRAMP v1.3 software (http://www.cuilab.cn/sramp). The YAP primers: forward, 5’-ACCCTCGTTTTGCCATGAAC-3’, reverse, 5’-TGTGCTGGGATTGATATTCCGTA-3’. The MAP4K4 primers: forward, 5’-GAAGAGACCCACCCAAGACG-3; reverse, 5’-TCGCTTTTTCTGTCCGCTTTC-3’.

RNA stability assay

To examine whether FTO modulates YAP and MAP4K4 mRNA stability, FTO-overexpressing or control cells were incubated with 5 μg/mL actinomycin D (Sigma, USA) at time points of 0, 0.5, 1, 2, and 4 h. Decay curves were generated by plotting the remaining transcript levels against time;changes in mRNA stability were assessed by comparing transcript degradation kinetics between experimental groups. Primer sequences used for RT-qPCR were identical to those described in the MeRIP-qPCR assay.

Protein extraction and western blot analysis

Total proteins were extracted from murine cardiac tissue or cultured H9c2 cells using RIPA lysis buffer (Beyotime Biotechnology). Following total protein quantification, 25 μg of protein per well was resolved on SDS-PAGE gels and then electrophoretically transferred onto nitrocellulose membranes. The membranes were blocked with 5% BSA prepared in TBS-T (Tris-buffered saline containing 0.1% Tween-20) for 1 h at room temperature and subsequently incubated overnight at 4 °C with primary antibodies against FTO (1:2000, Abcam, ab92821), YAP (1:1000, Proteintech, 13,584–1-AP), GPX4 (1:1000, Proteintech, 67,763–1-AP), FTH1 (1:1000, ABclonal, A19544), HO-1 (1:2000, Abcam, ab13248), MAP4K4 (1:1000, Proteintech, 19,288–1-AP), β-actin (1:2000, Proteintech, 66,009–1-Ig), and HRP-conjugated secondary antibody (1:3000, Proteintech, SA00001-0). Immunoreactive bands were imaged on a ChemiDoc luminescence detection system (Bio-Rad Laboratories).

Statistical analysis

Results were presented as mean ± standard deviation (SD), with experimental procedures independently replicated at least three biological replicates. Statistical comparison between two experimental cohorts was determined using an unpaired Student’s t-test, while comparisons involving more than two groups were evaluated by one-way ANOVA. All statistical computations were conducted using GraphPad Prism 9 (San Diego, CA, USA). P < 0.05 was regarded as statistically significant.

Results

DOX induces cardiac dysfunction and decreases FTO and YAP expression

To assess whether FTO protects against DIC by modulating YAP, we established an acute DIC mouse model. Echocardiographic analysis showed that LV-EF and LV-FS were significantly reduced 48 h after DOX treatment in mice (p < 0.01) (Fig. 1A). It was further confirmed that DOX induced pronounced myocardial injury, as evidenced by cardiomyocyte swelling, nuclear loss, degeneration, and necrosis observed in H&E-stained heart sections (Fig. 1B). TEM further showed that DOX induced severe ultrastructural abnormalities in cardiac tissues, including disrupted myofibrillar organization and mitochondrial damage, such as cristae loss, swelling, vacuolization, and membrane rupture (Fig. 1C). We subsequently evaluated the expression of FTO and YAP proteins in cardiac tissues. As shown in Fig. 1D, DOX administration markedly reduced both FTO and YAP proteins in mouse hearts (p < 0.05, DOX vs control). Thus, the FTO/YAP signaling axis may be involved in the pathogenesis of DIC.

Fig. 1.

FTO and YAP are decreased in DOX-induced myocardial injury. A Representative echocardiographic images and quantitative analysis of left ventricular ejection fraction (EF) and fractional shortening (FS) at 48 h after DOX administration (25 mg/kg). B Representative hematoxylin and eosin (H&E)-stained myocardial sections. Scale bar: 50 µm. C Transmission electron microscopy (TEM) images showing ultrastructural alterations in DOX-treated myocardial tissues. Scale bar: 2 µm. D Western blot analysis of YAP and FTO protein levels in DOX-treated myocardial tissues. Data are presented as mean ± SD; *P < 0.05, **P < 0.01 vs. control (n = 3)

FTO alleviates DIC and restores YAP expression in mice

To elucidate the regulations of FTO in DIC and YAP signaling, mice received tail-vein injections of AAV9-FTO, and FTO mRNA and protein expression were significantly enhanced in heart tissues at 4 weeks after AAV injection (Fig. 2B,C). Echocardiographic evaluation showed that AAV9-FTO significantly alleviated DOX-induced decreases of LV-EF and LV-FS (p < 0.01 and p < 0.05, respectively; Fig. 2D). Consistently, H&E staining showed that AAV9-mediated FTO overexpression markedly alleviated DOX-induced myocardial pathological injury, including cellular swelling, nuclear loss, degeneration, and necrosis (Fig. 2E).

Fig. 2.

FTO ameliorates DOX-induced myocardial injury and YAP inhibition. A Schematic of the experimental design in mice. B RT-qPCR analysis of FTO mRNA levels in myocardial tissues 4 weeks after AAV-FTO administration. C Western blot analysis of FTO protein in DOX-treated myocardial tissues 4 weeks post-AAV-FTO treatment. D Representative echocardiographic images and quantitative analysis of EF and FS at 48 h after DOX administration. E Representative H&E stained myocardial sections. Scale bar: 50 µm. F TEM images showing ultrastructural changes in DOX-treated myocardial tissues. Scale bar: 2 µm. G Intracellular Fe²⁺ levels in DOX-treated myocardial tissues. H MDA levels in DOX-treated myocardial tissues. I Western blot analysis of YAP and FTO protein levels in DOX-treated myocardial tissues. Data are presented as Mean ± SD. *P < 0.05, **P < 0.01 vs control; ^♯P < 0.05, ^♯♯P < 0.01 vs DOX (n = 3)

Because ferroptosis is a major mechanism underlying DIC, we next evaluated ferroptosis-related parameters in cardiac tissues, including mitochondrial ultrastructural changes, Fe²⁺, and MDA levels. TEM further confirmed that AAV-FTO attenuated DOX-induced ultrastructural damage, including myofibrillar disorganization, mitochondrial cristae loss, swelling, vacuolization, and membrane rupture (Fig. 2F). DOX significantly promoted Fe²⁺ and MDA accumulations in mouse hearts, which were markedly reversed by FTO overexpression (Fig. 2G, H). Furthermore, western blot analysis demonstrated that DOX markedly reduced YAP protein, which was significantly restored by AAV-FTO treatment in mouse hearts (p < 0.01, AAV-FTO + DOX vs. DOX; Fig. 2I).

FTO inhibits DOX-induced cardiac ferroptosis

To investigate whether FTO regulates cardiac ferroptosis during DIC in vitro, we first tested H9c2 cell viability under DOX treatment. CCK-8 analysis demonstrated that exposure to 200 nM DOX for 24 h resulted in an approximately 50% reduction in cell viability, which was selected for subsequent experiments (Fig. 3A). Next, stable H9c2 cells with either FTO overexpression or knockdown were generated and verified via analysis of FTO protein (p < 0.01; Fig. 3B). CCK-8 analysis demonstrated that DOX-induced decrease of H9c2 cell viability, which was markedly attenuated by FTO overexpression (p < 0.01, oeFTO + DOX vs DOX), but enhanced by FTO knockdown (p < 0.05, shFTO + DOX vs. DOX; Fig. 3C).

Fig. 3.

FTO inhibits DOX-induced cardiac ferroptosis. A Cell viability of H9c2 cells exposed to increasing concentrations of DOX to determine IC₅₀. B Validation of FTO-overexpressing (oeFTO) and FTO-knockdown (shFTO) H9c2 cell lines at the protein level. C Effect of FTO on H9c2 cell viability following DOX treatment, measured using CCK-8 assay after 24 h exposure to 200 nM DOX or PBS. D Reactive oxygen species (ROS) levels in control and FTO-transfected H9c2 cells following PBS or DOX treatment. Scale bar: 500 µm. E Intracellular Fe²⁺ levels in control and FTO-transfected H9c2 cells treated with or without DOX, assessed using FerroOrange staining. Scale bar: 50 µm. F Western blot analysis of GPX4, HO-1, and FTH expression, normalized to β-actin. Data presented as Mean ± SD. **P < 0.01 vs control; ^♯P < 0.05 vs DOX (n = 3)

To investigate the regulation of FTO on cardiac ferroptosis, we next assessed intracellular Fe²⁺ accumulation, ROS production, and ferroptosis-related proteins in H9c2 cells following DOX exposure. Fluorescence staining revealed that DOX significantly increased Fe²⁺ accumulation and ROS production, which were markedly attenuated by FTO overexpression in H9c2 cells (p < 0.01, DOX vs. NC; p < 0.05, oeFTO + DOX vs. DOX; Fig. 3D- E). Furthermore, western blot analysis showed that DOX treatment significantly reduced anti-ferroptotic proteins, GPX4, FTH, and HO-1 (p < 0.01 vs. Control), which were substantially restored by FTO overexpression (p < 0.05, oeFTO + DOX vs. DOX; Fig. 3F). Collectively, these findings showed that FTO alleviates DOX-induced cardiac ferroptosis.

YAP mediates the anti-ferroptotic effects of FTO

To investigate whether FTO mediates the anti-ferroptotic effects by regulating YAP, we assessed the role of FTO on YAP expression in DOX-treated H9c2 cells. As shown in Fig. 4A, DOX treatment significantly reduced YAP protein, which was markedly restored by FTO overexpression (p < 0.01, oeFTO + DOX vs. DOX). To further explore the role of YAP in FTO-mediated cardioprotection, YAP expression was silenced using si-YAP. As shown in Fig. 4B, YAP knockdown largely attenuated the protective effect of FTO overexpression against DOX-induced reduction in H9c2 cell viability. Moreover, FTO overexpression prominently rescued the expression of anti-ferroptotic markers (GPX4, FTH and HO-1) upon DOX stimulation (p < 0.05, oeFTO + DOX vs DOX), whereas YAP knockdown abolished such restorative effect (p < 0.05, DOX + oeFTO vs. DOX + si-YAP + oeFTO) (Fig. 4C).

Fig. 4.

YAP is essential for the anti-ferroptotic effects of FTO in DOX-treated H9c2 cells. A Western blot analysis of YAP and FTO protein levels in H9c2 cells under DOX treatment. B CCK-8 assay showing YAP’s role in mediating FTO’s protective effect on H9c2 cell viability under DOX exposure. C Western blot analysis of GPX4, HO-1, and FTH in normal or FTO-overexpressing H9c2 cells transfected with si-YAP and treated with 200 nM DOX for 24 h, normalized to β-actin. D Representative TEM images of mitochondrial morphology in control and FTO-overexpressing H9c2 cells treated with DOX. White, yellow, blue, and red arrows indicate normal mitochondria, increased membrane density, mitochondrial membrane rupture, and loss/disappearance of cristae, respectively. Scale bar: 250 nm. E MMP was measured using the JC-1 fluorescent probe in control and FTO-overexpressing H9c2 cells treated with DOX. Scale bar: 50 µm. Data presented as Mean ± SD, ***P < 0.001,**P < 0.01; *P < 0.05 vs DOX (n = 3)

We further evaluated mitochondrial morphology and function, as mitochondrial dysfunction is a key feature of ferroptosis. TEM revealed that DOX treatment induced severe mitochondrial damage, characterized by mitochondrial shrinkage, membrane thickening, and cristae collapse (Fig. 4D). These ultrastructural abnormalities were markedly attenuated by FTO overexpression, whereas YAP inhibition reversed the protective effects of FTO on mitochondrial integrity. MMP was assessed using JC-1 staining. DOX-treated cells predominantly exhibited green fluorescence, indicating MMP depolarization and mitochondrial dysfunction (Fig. 4E). In contrast, FTO overexpression increased red fluorescence intensity, suggesting improved MMP. However, these beneficial effects were significantly diminished following YAP silencing.

ATF4 is required for FTO/YAP-mediated anti-ferroptotic effects

Previous studies have reported that YAP suppresses ferroptosis by activating ATF4 (Gao et al. 2021). Thus, we further investigated whether the FTO/YAP axis protects against DOX-induced ferroptosis by regulating ATF4. As expected, DOX significantly downregulated the expression of ATF4 (DOX vs NC,p < 0.01), which was effectively rescued by FTO overexpression in H9c2 cells (oeFTO + DOX vs DOX, p < 0.05) (Fig. 5A). We further investigated whether YAP is involved in FTO-mediated regulation of ATF4 under DOX treatment. As shown in Fig. 5B, inhibition of YAP significantly blunted the FTO-induced upregulation of ATF4 in DOX-treated H9c2 cells (p < 0.05).

Fig. 5.

FTO/YAP inhibits DOX-induced cardiac ferroptosis by activating ATF4. A Western blot analysis of ATF4 expression in DOX-treated H9c2 cells under FTO overexpression. B Western blot showing the role of YAP in FTO-mediated regulation of ATF4 expression in DOX-treated H9c2 cells. C Western blot analysis of ferroptosis-related proteins (GPX4, HO-1, FTH) to examine ATF4’s role in FTO-mediated regulation in DOX-treated H9c2 cells. Data presented as Mean ± SD. **P < 0.01 vs control; ^♯P < 0.05 vs DOX; ^✛P < 0.05 vs DOX + oe FTO (n = 3)

To clarify whether ATF4 is required for the anti-ferroptotic effects of FTO, ATF4 was silenced using si-ATF4. As shown in Fig. 5C, FTO overexpression significantly rescued the DOX-induced downregulation of GPX4, FTH, and HO-1; however, this effect was largely attenuated following ATF4 knockdown in H9c2 cells (p < 0.05).

MAP4K4 inhibition potentiates the anti-ferroptotic effects of the FTO/YAP/ATF4 axis

Interestingly, FTO was also reported to enhance MAP4K4 expression, which could inhibit YAP signaling (Heallen and Martin 2023). We therefore explored whether MAP4K4 mediates the regulation of FTO on YAP signaling and DOX-induced cardiac ferroptosis. As shown in Fig.6A, DOX exposure markedly decreased MAP4K4 expression, whereas FTO overexpression markedly attenuated the DOX-induced decrease of MAP4K4. We next examined the influence of the MAP4K4 inhibitor DMX-5804 at different concentrations on MAP4K4 expression and cell viability in H9c2 cells. As shown in Fig. 6B-C, 10 μM DMX-5804 effectively suppressed MAP4K4 expression without affecting cell viability, whereas 20 μM DMX-5804 significantly reduced cell viability.

Fig. 6.

MAP4K4 inhibition enhances the anti-ferroptosis effect of FTO/YAP/ATF4. A Western blot analysis of MAP4K4 protein expression in control and FTO-overexpressing H9c2 cells treated with 200 nM DOX for 24 h. B Effect of different concentrations of the MAP4K4 inhibitor DMX-5804 on MAP4K4 expression in H9c2 cells. C Immunofluorescence staining of YAP (green) and DAPI (blue) in DOX-treated H9c2 cells. Scale bar: 10 μm. D Western blot analysis showing MAP4K4’s role in FTO-mediated regulation of YAP and ferroptosis-related proteins (GPX4, HO-1, FTH). Data presented as Mean ± SD. P < 0.05, **P < 0.01, (n = 3)

Given that MAP4K4 promotes YAP phosphorylation, leading to its cytoplasmic degradation and reduced nuclear translocation, we next examined whether DMX-5804 affects YAP nuclear localization in DOX-treated H9c2 cells. As shown in Fig. 6D, DOX reduced YAP nuclear translocation, while DMX-5804 promoted YAP nuclear translocation in H9c2 cells. Moreover, whether MAP4K4 mediates the regulation of FTO on ferroptosis was further investigated. Notably, pharmacological inhibition of MAP4K4 further enhanced FTO-mediated upregulations of YAP, GPX4, FTH, and HO-1 in DOX-treated H9c2 cells (Fig. 6E).

FTO up-regulates MAP4K4 and YAP through m⁶A demethylation

Accumulating evidence has shown that m⁶A modification participates in the regulation of MAP4K4 and YAP (Ke et al. 2022; Kitakata et al. 2022). To determine whether FTO regulates MAP4K4 and YAP expression through m⁶A demethylation under DOX treatment, we first quantified total RNA m⁶A levels in H9c2 cells. As shown in Fig.7A, DOX treatment markedly increased total RNA m⁶A level, which was significantly reduced by FTO overexpression, indicating that FTO counteracts DOX-induced m⁶A hypermethylation. Bioinformatic prediction using SRAMP v1.3 revealed that MAP4K4 and YAP transcripts harbor multiple putative high-confidence m⁶A modification sites (Fig. 7B). To further assess whether FTO directly modulates m⁶A methylation on MAP4K4 and YAP transcripts, we performed MeRIP and qRT-PCR analysis. MeRIP-qPCR analysis showed that FTO overexpression significantly reduced m⁶A enrichment of MAP4K4 and YAP mRNAs in DOX-treated H9c2 cells (Fig. 7C, p < 0.05). We next assayed the influence of FTO on MAP4K4 and YAP mRNA stability using actinomycin D-mediated transcriptional inhibition. As shown in Fig. 7D, FTO-overexpressing H9c2 cells exhibited significantly higher residual MAP4K4 and YAP mRNA levels at 2 h and 4 h after actinomycin D treatment (p < 0.05), suggesting enhanced transcript stability. Collectively, these findings demonstrate that FTO promotes the stability of MAP4K4 and YAP mRNAs by reducing their m⁶A methylation, thereby contributing to the regulation of the FTO/MAP4K4/YAP signaling axis during DIC.

Fig. 7.

FTO mediates m⁶A methylation of MAP4K4 and YAP. A Total RNA m⁶A levels in control and FTO-overexpressing H9c2 cells treated with 200 nM DOX for 24 h. B SRAMP bioinformatic prediction of putative m⁶A modification sites within MAP4K4 and YAP mRNA. C FTO-mediated m⁶A modification of MAP4K4 and YAP mRNA in DOX-treated H9c2 cells, determined by MeRIP-qPCR. D Effect of FTO on MAP4K4 and YAP mRNA stability. Degradation curves were generated using qRT-PCR after actinomycin D treatment at the indicated time points. Data are presented as mean ± SD; * P < 0.05, **P < 0.01, (n = 3)

Discussion

Ferroptosis is a major contributor to the pathology of DIC. DOX and its metabolites disrupt intracellular iron homeostasis (Ren et al. 2023), causing iron accumulation, lipid peroxidation, and mitochondria-dependent cardiac ferroptosis (Schirone et al. 2023). Our data demonstrated that FTO and YAP expression were markedly downregulated in hearts and H9c2 cells under DOX exposure. Importantly, FTO was found to attenuate DOX-induced cardiac ferroptosis by enhancing YAP/ATF4 signaling. Notably, FTO also increased MAP4K4 expression, whereas pharmacological inhibition of MAP4K4 further potentiated the anti-ferroptotic effects of the FTO/YAP/ATF4 axis during DIC.

Emerging studies have revealed that m⁶A RNA modification is a key regulatory mechanism governing ferroptosis in DIC and YAP signaling. Previous studies have shown that inhibition of METTL3 or METTL14, as well as overexpression of FTO, alleviates DIC by suppressing ferroptosis (Ke et al. 2022; Kitakata et al. 2022; Li et al. 2025). Previous studies have shown that circ-ZNF609 attenuates DIC by upregulating FTO expression (Shao et al. 2014). Circ-ZNF609 was also found to inhibit DIC by improving YAP mRNA stability through interactions with YTHDF1 and YTHDF2 (Shao et al. 2025). We also found that FTO inhibited cardiac ferroptosis in DIC via m⁶A demethylation of YAP. Consistently, FTO-mediated m⁶A demethylation enhances YAP mRNA stability and protects against I/R-induced myocardial ferroptosis (Matsuda et al. 2016). Similar to FTO, ALKBH5, another m⁶A demethylase, has also been reported to up-regulate YAP expression by modulating m⁶A methylation and YTHDF-mediated RNA stability (Sun et al. 2022; Tadokoro et al. 2020). Accumulating evidence highlights YAP as a central regulator of both cardiovascular pathophysiology and ferroptosis resistance. For example, restoration of YAP expression alleviates heart failure progression (Tang et al. 2025; Wang et al. 2021), whereas YAP inhibition exacerbates I/R-induced murine myocardial injury (Wang et al. 2022). YAP significantly improves cardiac function, attenuates fibrosis, and suppresses hypertrophic remodeling following myocardial injury (Wang et al. 2024). Similarly, DOX treatment markedly suppresses YAP expression in cardiomyocytes (Wang et al. 2025), and reduced YAP/TAZ correlates with the severity of DOX-induced cardiomyocyte injury (Yang et al. 2024). Notably, YAP is critical for melatonin-induced cardioprotection against DOX-triggered ferroptosis (Gao et al. 2021).

Mechanistically, YAP promotes the transcription of anti-ferroptosis genes, including ATF4, NEDD4L, and PITX2, thereby suppressing ROS accumulation and ferroptosis by interacting with TEAD transcription factors (Heallen and Martin 2023; Identification 2001; Yang et al. 2024). In this study, we found that FTO upregulated ATF4 expression through YAP, and that ATF4 was essential for the anti-ferroptotic effect of FTO. Notably, YAP inhibition abolished FTO-induced upregulation of ATF4 in DIC. ATF4, a well-established stress-responsive transcription factor, has been reported to inhibit sorafenib-induced cardiac ferroptosis(Yin et al. 2024). It is favored that estrogen-responsive B box protein exerts cardioprotective effects against DOX-induced ferroptosis through GRP78-dependent ATF4/Nrf2 activation (Yu et al. 2023). Consistently, ATF4 has also been reported to cooperate with Nrf2 in regulating antioxidant and anti-ferroptotic genes, including xCT and HO-1 (Zhang et al. 2025; Zhao et al. 2024). In addition, ATF4 induces HSPA5 expression, which further suppresses ferroptosis by stabilizing GPX4 (Zhou et al. 2023). Our data also confirmed that silencing ATF4 partially reversed the FTO/YAP-induced upregulation of xCT, HO-1, and GPX4.

Notably, MAP4K4, an upstream regulator of YAP phosphorylation, has also been reported to promote DIC (Meng et al. 2023). Pharmacological inhibition of MAP4K4 also attenuates cardiomyocyte apoptosis and reduces infarct size by restoring mitochondrial function (Ndubaku et al. 2015). MAP4K4 has been identified to promote YAP phosphorylation and cytoplasmic degradation by activating LATS1/2 kinases (Zhu et al. 2017). It was favored that MAP4K4 suppressed oncogenic activity by up-regulating LATS2 and SASH1 within the noncanonical Hippo pathway (Zhuang et al. 2023). Our results demonstrated that DOX-induced cytotoxicity was associated with decreased MAP4K4 expression in cardiomyocytes. As MAP4K4 negatively regulates YAP activity, this reduction may represent an adaptive compensatory response to sustain YAP signaling under DOX stress. Interestingly, we found that FTO overexpression not only restored YAP expression but also concurrently increased MAP4K4 levels through m⁶A-dependent demethylation. MAP4K4 inhibitor DMX-5804 markedly promoted YAP nuclear translocation and the anti-ferroptotic effects mediated by the FTO/YAP/ATF4 signaling. Consistently, FTO-mediated m⁶A demethylation has been reported to upregulate MAP4K4 expression and promote gastric cancer progression (Qiu et al. 2023).

Limitations

There were several limitations in this study. First, H9c2 cells, derived from embryonic rat cardiomyocytes, are commonly used to investigate DIC, they cannot fully replicate the physiological and metabolic characteristics of mature cardiomyocytes. Second, this study utilized a short-term acute DOX-induced myocardial injury model with a relatively small sample size, and AAV-FTO was administered prior to DOX exposure as a preventive strategy. Moreover, the combination of FTO activation and MAP4K4 inhibition may raise concerns regarding potential off-target effects. Thus, further studies with larger cohorts and comprehensive safety assessments are required. Third, the direct regulatory mechanism between YAP and ATF4 was not fully clarified in the present study. Fourth, DOX-induced cardiotoxicity is mediated by apoptosis, necroptosis, and pyroptosis, which were not comprehensively assessed in this study. Fifth, we detected FTO protein in the heart, but did not verify ectopic expression in other organs, such as the liver and kidney, to eliminate systemic confounding factors. Fifth, FTO protein expression was confirmed in the heart, but its expression in peripheral organs, including the liver and kidney, was not evaluated, thereby not fully excluding systemic confounding effects. Finally, although MeRIP-qPCR and mRNA stability analyses indicated that FTO mediated m⁶A demethylation of MAP4K4 and YAP transcripts, direct interactions between FTO and these mRNAs remain to be further validated.

Conclusion

FTO markedly attenuates DOX-induced cardiac ferroptosis by promoting YAP/ATF4 signaling through m⁶A demethylation of YAP transcripts. Meanwhile, FTO also enhances MAP4K4 expression via m⁶A demethylation, and pharmacological inhibition of MAP4K4 further potentiates the anti-ferroptotic effects of FTO/YAP/ATF4 signaling against DOX-induced cardiotoxicity.

Abbreviations

AAV9

Adeno-associated virus serotype 9

ATF4

Activating transcription factor 4

CCK-8

Cell Counting Kit-8

cTnT

Cardiac troponin T

DOX

Doxorubicin

FTO

Fat mass and obesity-associated protein

FTH

Ferritin heavy chain

GPX4

Glutathione peroxidase 4

H&E

Hematoxylin and eosin

HO-1

Heme oxygenase-1

I/R

Ischemia–reperfusion

LATS

Large tumor suppressor kinase

LV-EF

Left ventricular ejection fraction

LV-FS

Left ventricular fractional shortening

m⁶A

N6-methyladenosine

MAP4K4

Mitogen-activated protein kinase kinase kinase kinase 4

MDA

Malondialdehyde

MeRIP

M⁶A RNA Immunoprecipitation

METTL3

Methyltransferase-like 3

METTL14

Methyltransferase-like 14

MMP

Mitochondrial membrane potential

NC

Negative control

NRF2

Nuclear factor erythroid 2–related factor 2

ROS

Reactive oxygen species

TEM

Transmission electron microscopy

TEAD

TEA domain transcription factor

YAP

Yes-associated protein

Author contributions

Conceptualization: Junnan Wang. Methodology:Dandan Song. Investigation: Youyou Yan, Dandan Song, and Longbo li. Writing—original draft: Youyou Yan. Writing—review and editing: Junnan Wang. Conceptualization: Junnan Wang. Methodology:Dandan Song. Investigation: Youyou Yan, Dandan Song, and Longbo li. Writing—original draft: Youyou Yan. Writing—review and editing: Junnan Wang.

Funding

This study was supported by the Department of Science and Technology of Jilin Province (Grant No. 20240305031YY), the World Medical Science and Technology Special Project (WFCM2024010), the Medical and Health Talent Project of Jilin Province (2025wszx-Jc26), the Project of “Medical + X″ Interdisciplinary Innovation Team of Norman Bethune Health Science Center of Jilin University (2025JBGS03) and the Scientific and Technological Innovation Project of Changbaishan Laboratory, Jilin Province (CBS2026009-05).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

No datasets were generated or analysed during the current study.

Declarations

Ethical Approval

All animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee of Jilin University (2023227).

Conflict of interest

The authors declare no competing interests.