All experimental procedures involving animals were undertaken in accordance with the National Institutes of Health guidelines and received approval from the Animal Care and Use Committee of Jiaxing University Medical College (Approval No. JUMC2021-095). Wild type (WT) C57BL/6 mice (6–8 weeks old; Cyagen, Suzhou, Jiangsu, China) and Anxa1 knockout (KO) mice with a C57BL/6 background (6–8 weeks old; Cyagen, Suzhou, Jiangsu, China) were maintained in a specific pathogen-free facility under controlled conditions: relative humidity 40–70%, ambient temperature of 20–25 °C, 12-h light/dark cycle, and free access to food/water. Euthanasia was induced by placing animals in a chamber pre-filled with 100% medical-grade CO₂ at a flow rate displacing 30–70% of the chamber volume per minute. After $\ge$5 min post-respiratory arrest, cervical dislocation was performed at the C1–C2 vertebrae to confirm death. To eliminate potential confounding effects of female sex hormones, only male mice were utilized. To minimize bias, a randomized allocation protocol was implemented. Mice were randomly assigned to either experimental or control cohorts. Cardiac I/R injury modeling and echocardiographic assessments were conducted by an operator blinded to group assignments. Group allocation information was concealed from the operator throughout both the experimental execution and subsequent data analysis phases, ensuring blinding integrity.

Following standard procedures, mice underwent a transient left anterior descending (LAD) coronary artery occlusion to induce cardiac I/R injury. Under general anesthesia (2% isoflurane), mice underwent endotracheal intubation with mechanical ventilation support. A 5-mm left thoracotomy between the 4th and 5th intercostal spaces was performed to expose the cardiac apex. A silk suture (8-0) was positioned 2–3 mm distal to the LAD origin, adjacent to the junction of the left atrial appendage and pulmonary artery. Coronary occlusion was achieved by tightening the suture loop, with successful ischemia induction confirmed by ST-segment elevation on continuous electrocardiogram monitoring and pallor of the distal myocardium. Following 30 min of ischemia, reperfusion (2 h) was initiated by suture release, evidenced by myocardial blush restoration. Postoperatively, mice were allowed to recover from anesthesia on a warm pad with ad libitum access to water and moistened chow. Sham-operated mice underwent identical procedures except for coronary occlusion (no LAD ligation).

WT mice were randomized into 3 groups in part I (sham, MI + adnull, and MI + adFto; n = 6). WT and Anxa1 KO mice were respectively randomized into 3 groups in part II (sham, MI + adnull, and MI + adFto; n = 6). Fto overexpression was achieved in MI mice via intramyocardial injection of a recombinant adeno-associated virus serotype 9 vector (AVV9) encoding Fto (adFto). Both adFto and a control vector (adnull) were purchased from Genechem (Shanghai, China). To ensure sufficient protein overexpression during cardiac I/R injury, intramyocardial delivery of ad vectors (adFto/adnull, 1 $\times$ 10⁹ vg/mouse) was performed via left ventricular multi-point injection. Using a 30-gauge ultrafine needle coupled to an insulin syringe, a total of 20 µL AAV9 was delivered as four 5 µL intramyocardial injections into the left-ventricular anterior wall bordering the ischemic area, from apex to base, using a 30-gauge Hamilton syringe, with approximately 60 s per site [20].

Cardiac functional evaluation was performed using a transthoracic echocardiography (RMV-707B; VisualSonics, Toronto, ON, Canada) with a Vevo 2100 high-resolution imaging system (FujiFilm VisualSonics, Inc., Toronto, ON, Canada) equipped with a 40-MHz single-element transducer. Following anesthesia induction via 2% isoflurane, mice were maintained under 0.8–1.5% vaporized isoflurane through a gas anesthesia system. M-mode acquisitions were obtained from parasternal long-axis views at papillary muscle level, with 3 consecutive cardiac cycles recorded for offline analysis. Left ventricular functional indices, including ejection fraction (EF) and fractional shortening (FS).

Serum samples were obtained through retro-orbital venous plexus puncture under isoflurane anesthesia (2%–3% for induction and 1%–2% for maintenance). Whole blood was processed by centrifugation (3000 $\times$g, 15 min, 4 °C) to isolate serum supernatant. Cardiac injury biomarkers, including Troponin T (TnT) and creatine kinase-myocardial band isoenzyme (CK-MB) were quantified using species-specific ELISA kits (#MBS761403/#MBS2701855; MyBiosource, San Diego, CA, USA) following standardized protocols.

Total RNA isolation from heart tissues was conducted with TRIeasyTM total RNA extraction reagent (#10606ES60; Yeasen, Shanghai, China) following standard protocols. First-strand complementary DNA (cDNA) synthesis was subsequently performed using Hifair® II 1st strand cDNA synthesis kit (#11119ES60; Yeasen, Shanghai, China). Quantitative PCR amplification was achieved with Hieff® qPCR SYBR green master mix (high Rox plus; #11203ES08; Yeasen). Gene expression was quantified using the 2^{-Δ⁢Δ⁢CT} method, normalized to glyceraldehyde-3-phosphate dehydrogenase (Gapdh) expression [21]. Amplification by quantitative PCR employed the subsequent primer sets: Gapdh, Forward-5^′-CATCACTGCCACCCAGAAGACTG-3^′, Reverse-5^′-ATGCCAGTGAGCTTCCCGTTCAG-3^′; Anxa1, Forward-5^′-TGTATCCTCGGATGTTGCTGCC-3^′, Reverse-5^′- CCATTCTCCTGTAAGTACGCGG-3^′.

Cardiac tissue homogenization was performed in radioimmune precipitation assay buffer (#G2002; Servicebio, Wuhan, Hubei, China) supplemented with phenylmethanesulfonyl fluoride (1 mM; #G2008; Servicebio, Wuhan, Hubei, China). Prior to electrophoresis, the protein concentration in each sample was measured with a commercial bicinchoninic acid assay kit (#G2026; Servicebio, Wuhan, Hubei, China). Equal amounts of lysates were resolved through 10–12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and electrophoretically transferred onto polyvinylidene fluoride membranes, followed by blocking with 5% skim milk/Tris-buffered saline with 0.1% Tween-20. The membranes were immuno-detected with primary antibodies against Fto, Anxa1, B cell lymphoma-2 (Bcl-2), Bcl-2-associated X protein (Bax), apoptosis-associated speck-like protein (ASC), cleaved-Caspase 1 (cle-Caspase 1), pro-Caspase 1, Nlrp3, gasdermin D-N-terminal domain (N-Gsdmd), gasdermin D (Gsdmd), and Gapdh. Information on all antibodies is displayed in Supplementary Table 1. After TBST washing, the membranes were probed with a secondary antibody at ambient temperature for 60 min. Protein signals were developed using an enhanced ECL chemiluminescent substrate kit (#36222ES76; Yeasen). Immunoreactive bands were quantified densitometrically using ImageJ software (1.54p; NIH, Bethesda, MD, USA), with Gapdh serving as the loading control.

The cardiac tissues were cut into small pieces, homogenized, and centrifuged to get the supernatants (5 min, 5000 $\times$g). The harvested supernatants were employed for the detection of pro-inflammatory cytokines tumor necrosis factor alpha (TNF-$\alpha$), IL-6, and IL-1$\beta$. All detection steps were performed in strict accordance with the instructions provided with the ELISA kit (#MBS2124239/#MBS450807/#MBS2021142; MyBiosource, San Diego, CA, USA).

Cardiac tissues were harvested and fixed in 4% paraformaldehyde (#E672002; Sangon, Shanghai, China), followed by embedding into paraffin and sectioning at 4 µm intervals. Subsequently, the sections underwent deparaffinization, dehydration via xylene, and ethanol gradient immersion, followed by Masson’s trichrome staining according to standard procedures. Collagen fiber content was analyzed according to the following equation: Collagen fiber (%) = (blue area pixels/total area pixels) $\times$ 100%.

Total RNA samples prepared with TRIzol extraction reagent were subjected to m6A quantification via the m6A RNA methylation quantification kit (#P-9005; Epigentek, Flushing, NY, USA). RNA aliquots (200 ng/sample) were sequentially incubated with capture antibody, detection antibody, enhancer solution, and chromogenic substrate in 96-well plates. Absorbance values at 450 nm were measured with a microplate reader (Bio-Rad, Hercules, CA, USA), where experimental values were derived from established standard curve calculations.

Total RNA from left ventricles was extracted with TRIzol and treated with DNase I. m6A RNA immunoprecipitation (IP) was performed using an anti-m6A antibody with species-matched IgG as a negative control according to the manufacturer’s protocol. Information on all antibodies is displayed in Supplementary Table 1. After IP, RNA was purified, reverse-transcribed, and qPCR was carried out for Anxa1. Enrichment was quantified as % input and as fold over IgG.

Paraffin-embedded cardiac tissues were sectioned into 4 µm sections for IHC analysis with a 3,3-diaminobenzidine (DAB) detection system. Endogenous peroxidase was inactivated by sequential deparaffinization, rehydration, and 15-min incubation in 3% H₂O₂-methanol solution, followed by 20-min antigen retrieval at 95 °C. Tissue sections were treated overnight at 4 °C with anti-Nlrp3 primary antibody (#D261589; at 1:25 dilution; Sangon, Shanghai, China), followed by 30-min incubation with a secondary antibody (#D110073; at 1:1000 dilution; Sangon, Shanghai, China). Reaction products were visualized after counterstaining with the ultra-sensitive horseradish the catalase DAB color kit (#C510023; Sangon, Shanghai, China). Images were captured at 100$\times$ magnification using a microscope (Olympus, Tokyo, Japan).

Experimental results are presented as mean $\pm$ standard error of mean (SEM) (minimum 3 repeats). Analyses were conducted in GraphPad Prism 9 (GraphPad Software, San Diego, CA, USA), with multi-group comparisons meeting normality and homogeneity of variance assumptions analyzed by one-/two-way analysis of variance (ANOVA) and Tukey’s multiple comparisons test. Statistical significance was defined at p 0.05.

To verify that Fto-mediated m6A demethylation of Anxa1 mitigates cardiac I/R injury in vivo, we injected adFto or adnull intramyocardially in mice, followed by performing the MI surgery and reperfusion (Fig. 1A). Changes in Fto protein levels were detected in heart samples. As expected, MI mouse-derived heart samples possessed lower Fto protein levels than those from sham mice, and administration of adFto increased Fto protein levels in heart samples from MI mice (Fig. 1B). Cardiac function was assessed by Echo, as displayed in Fig. 1C. We observed significant reductions in both EF (56.17% vs. 34.00%, p 0.001) and FS (35.67% vs. 12.67%, p 0.001) values in the hearts of mice subjected to MI, but Fto overexpression improved these reductions (EF: 34.00% vs. 46.67%, p 0.001; FS: 12.67% vs. 20.67%, p 0.001) (Fig. 1D,E). The myocardial injury parameters CK-MB and TnT in the serum of MI mice were significantly higher than those of sham mice (CK-MB: 149.7 mg/mL vs. 902.5 mg/mL, p 0.001; TnT: 2.34 pg/mL vs. 15.02 pg/mL, p 0.001), yet ectopic expression of Fto ameliorated serum CK-MB (902.5 mg/mL vs. 606.6 mg/mL, p 0.001) and TnT (15.02 pg/mL vs. 12.63 pg/mL, p 0.001) levels for MI mice (Fig. 1F,G). Masson’s staining showed a higher content of collagen fibers in MI mouse-derived heart tissues (19.43% vs. 38.80%, p 0.001), but exogenous overexpression of Fto decreased collagen fiber content in heart tissues (38.80% vs. 25.84%, p 0.001) (Fig. 1H,I). Considering that the main role of Fto is demethylation modification, we tested m6A levels in heart tissues. Total m6A levels were significantly up-regulated in heart tissues of MI mice (0.02818% vs. 0.04672%, p 0.001), while exogenous expression of Fto reversed the increase in m6A levels in heart tissues (0.04672% vs. 0.03084%, p 0.001) (Fig. 1J). Moreover, m6A-RIP-qPCR showed that m6A enrichment on Anxa1 decreased with Fto overexpression (p 0.001) (Fig. 1K). IgG controls remained low across groups. These data support Fto-dependent transcript-level demethylation of Anxa1 in vivo. Additionally, a significant reduction in mRNA and protein levels of Anxa1 was observed in heart tissues of mice undergoing MI (p 0.001 and p 0.001), whereas exogenous expression of Fto potently raised Anxa1 mRNA and protein levels (p 0.001) (Fig. 1L,M). Collectively, these results indicate that Fto improved cardiac function and reduced injury and fibrosis in MI hearts, accompanied by reduced m6A enrichment on Anxa1 and increased Anxa1 expression.

Fig. 1.

Fto overexpression alleviates cardiac I/R injury and enhances Anxa1 expression in mouse models. (A) Experimental strategy in mouse models with MI. Mice underwent LAD occlusion for 30 min after intramyocardial injection of adFto or adnull (1 $\times$ 10⁹ vg/mouse) for 3 weeks, followed by reperfusion for 2 h. After four weeks, the mice underwent Echo, followed by collecting blood and heart samples to analyze cardiac I/R injury. (B) Western blot was carried out to detect Fto protein levels in heart samples (n = 3). (C–E) Representative Echo images for mice in each group and cardiac function indices EF (%) and FS (%) (n = 6). (F,G) Serum CK-MB and TnT levels were analyzed by ELISA (n = 6). (H,I) Representative images and a quantitative histogram of Masson’s staining on heart sections (n = 6). Scale bar: 500 µm. (J) Total m6A levels in heart tissues were assessed with an anti-m6A antibody and colorimetric quantification (n = 3). (K) m6A-RIP-qPCR was conducted to analyze the enrichment m6A on Anxa1 in heart tissues (n = 3). (L,M) Relative mRNA and protein levels of Anxa1 in heart tissues were detected by RT-qPCR and western blot (n = 3). Data are shown as mean $\pm$ SEM. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test in (B,D,E, and F–M); ***p 0.001. Fto, fat mass and obesity-associated protein; I/R, ischemia-reperfusion; Anxa1, Annexin A1; MI, myocardial infarction; LAD, left anterior descending; adFto, adenovirus encoding Fto; adnull, a control vector for adFto; Echo, echocardiography; CK-MB, creatine kinase-myocardial; TnT, troponin T; ELISA, enzyme-linked immunosorbent assay; m6A, N6-methyladenosine; EF, ejection fraction; FS, fractional shortening; RIP, RNA immunoprecipitation; qPCR, quantitative polymerase chain reaction; SEM, standard error of the mean; ANOVA, analysis of variance; Gapdh, glyceraldehyde-3-phosphate dehydrogenase.

To identify that Fto exerts cardioprotective effects by up-regulating Anxa1, we constructed MI models after adFto or adnull administration using WT or Anxa1 KO mice, followed by reperfusion. Briefly, a 2 $\times$ 2 factorial analysis within MI cohorts (genotype: WT vs Anxa1 KO; treatment: adnull vs. adFto) using two-way ANOVA with a genotype $\times$ treatment interaction. Sham values are provided to illustrate baseline ranges and confirm I/R effects. At the experimental endpoint, Anxa1 protein could not be detected in Anxa1 KO mouse-derived cardiac tissues, and there were no significant differential alterations in Fto protein levels between heart tissues from WT and Anxa1 KO mice (Fig. 2A). We discovered that Anxa1 knockout didn’t affect EF values significantly in normal mice, but decreased FS values slightly. Like WT mice, Anxa1 KO mice subjected to MI showed a significant reduction in EF (61.50% vs. 25.17%, p 0.001) and FS (29.67% vs. 15.17, p 0.001) values. Interestingly, Fto overexpression ameliorated the reduced EF and FS values in WT mice undergoing MI but did not work in Anxa1 KO mice (EF: 25.17% vs. 24.83, p $>$ 0.9999; 15.17% vs. 15.00%, p $>$ 0.9999) (Fig. 2B–D). The genotype $\times$ treatment interaction was significant for EF (F [2, 30] = 25.41, p 0.001) and FS (F [2, 30] = 38.14, p = 0.0130), indicating that the functional benefit of adFto depends on Anxa1. Furthermore, adFto reduced CK-MB and cTnT in WT but not in Anxa1-KO (p $>$ 0.9999), with a significant genotype $\times$ treatment interaction (CK-MB: F [2, 30] = 424.1, p 0.001; cTnT: F [2, 30] = 14.79, p 0.001) (Fig. 2E,F). In addition, heart sections of Anxa1 KO mice presented a substantial elevation in collagen fibers post-MI, but Fto overexpression did not affect changes in cardiac collagen fibers following Anxa1 knockout (36.76% vs. 35.15%, p = 0.2301) (interaction: F [2, 30] = 211.0, p 0.001) (Fig. 2G,H). Together, these results suggested that Fto protects against cardiac I/R injury by increasing Anxa1 expression.

Fig. 2.

Fto mediates Anxa1 expression against cardiac I/R injury. After intramyocardial injection of adFto or adnull (1 $\times$ 10⁹ vg/mouse) for 3 weeks, both WT and Anxa1 KO mice underwent sham or LAD occlusion for 30 min, followed by reperfusion for 2 h. (A) Protein levels of Fto and Anxa1 in heart samples from different groups were detected by western blot (n = 3). (B–D) Representative Echo images for mice in each group and histograms of cardiac function parameters EF (%) and FS (%) (n = 6). (E,F) Measurement of serum CK-MB and TnT levels was performed with ELISA (n = 6). (G,H) Representative images of Masson’s staining with heart sections and a quantitative histogram for collagen fiber content (n = 6). Scale bar: 500 µm. Data are shown as mean $\pm$ SEM. Statistical analysis was performed using two-way ANOVA followed by Tukey’s post hoc test in (A, C–F, and H); *p 0.05, **p 0.01, and ***p 0.001. Fto, fat mass and obesity-associated protein; I/R, ischemia-reperfusion; Anxa1, Annexin A1; WT, wild type; KO, Anxa1 knockout; LAD, left anterior descending; adFto, adenovirus encoding Fto; adnull, a control vector for adFto; Echo, echocardiography; CK-MB, creatine kinase-myocardial; TnT, troponin T; ELISA, enzyme-linked immunosorbent assay; EF, ejection fraction; FS, fractional shortening; SEM, standard error of the mean; ANOVA, analysis of variance.

Given the pathological significance of apoptosis and inflammation in cardiac I/R injury after MI, we proceeded to analyze whether Fto exerts its effects on cardiac tissue apoptosis and inflammation via Anxa1. The results showed that cardiac samples from WT and Anxa1 KO mice exhibited higher Bax protein levels and lower Bcl-2 protein levels after MI (p 0.001). Fto overexpression partly reversed Bax protein levels in WT and Anxa1 KO mice (p 0.001), suggesting that Fto regulates Bax protein levels by mediating Anxa1 and other genes. However, in Anxa1 KO mice, the effect of adFto on the Bcl-2 restoration was absent (p = 0.2561) (interactions: F [2, 30] = 49.65, p 0.001), suggesting that Fto regulates Bax protein levels primarily through mediating Anxa1 expression (Fig. 3A). As for the inflammatory response, cardiac samples of WT and Anxa1 KO mice experiencing MI presented higher levels of TNF-$\alpha$, IL-6, and IL-1$\beta$ (p 0.001), whereas Fto overexpression attenuated the levels of these inflammatory cytokines in WT mice (p 0.001) but not in Anxa1 KO mice (p = 0.8749 and p $>$ 0.9999), yielding significant interactions (TNF-$\alpha$: F [2, 30] = 1127, p 0.001; IL-6: F [2, 30] = 11798, p 0.001; IL-1$\beta$: F [2, 30] = 9273, p 0.001), which suggested that Fto modulates inflammatory cytokines mainly through mediating Anxa1 expression (Fig. 3B–D). Collectively, apoptotic and inflammatory readouts were attenuated by adFto in WT but not in Anxa1 KO, in association with higher Anxa1 expression, indicating that these benefits are linked to Anxa1.

Fig. 3.

Fto reduces cardiac apoptosis and inflammation via elevating Anxa1 expression post-MI. (A) Protein levels of Bax and Bcl-2 in heart samples were detected by western blot (n = 3). (B–D) The levels of inflammatory cytokines (TNF-$\alpha$, IL-6, and IL-1$\beta$) in heart samples were detected by ELISA (n = 6). Data are shown as mean $\pm$ SEM. Statistical analysis was performed using two-way ANOVA followed by Tukey’s post hoc test in (A–D); ***p 0.001. Fto, fat mass and obesity-associated protein; Anxa1, Annexin A1; Bcl-2, B cell lymphoma-2; Bax, Bcl-2-associated X protein; TNF-$\alpha$, tumor necrosis factor alpha; IL, interleukin; ELISA, enzyme-linked immunosorbent assay; EF, ejection fraction; FS, fractional shortening; SEM, standard error of the mean; ANOVA, analysis of variance.

To verify the effect of the Fto/Anxa1 axis on cardiac Nlrp3 activation and pyroptosis, we detected NLRP3 protein levels by IHC staining. As shown in Fig. 4A, Nlrp3 protein levels were higher in WT and Anxa1 KO mouse-derived heart samples post-MI, yet Fto overexpression reduced Nlrp3 protein levels in WT mice but not in Anxa1 KO mice. Additionally, higher protein levels of ASC, cleaved caspase-1/pro caspase-1 ratio, Nlrp3, and N-Gsdmd/Gsdmd were observed in heart samples of WT and Anxa1 KO mice with MI (p 0.001), whereas Fto up-regulation partially overturned the alteration of these proteins in the hearts of WT mice (p = 0.0064 and p 0.001) but not Anxa1 KO mice for cleaved caspase-1/pro caspase-1 ratio and N-Gsdmd/Gsdmd protein levels (p = 0.1346 and p = 0.8306) (ASC: F [2, 30] = 7.578, p = 0.0074; cleaved caspase-1/pro caspase-1 ratio: F [2, 30] = 56.30, p 0.001; Nlrp3: F [2, 30] = 11.54, p = 0.0016; N-Gsdmd/Gsdmd: F [2, 30] = 292.9, p 0.001;) (Fig. 4B). Altogether, these outcomes indicated that suppression of the Nlrp3 axis and pyroptosis happens in conjunction with Anxa1 upregulation downstream of Fto.

Fig. 4.

Fto represses Nlrp3-mediated pyroptosis in the heart following MI through Anxa1. (A) IHC staining was conducted to detect Nlrp3 protein levels in heart samples from different groups. Scale bar: 100 µm. (B) Protein levels of ASC, cleaved caspase-1/pro caspase-1 ratio, Nlrp3, and N-Gsdmd/Gsdmd in heart samples were evaluated by western blot. Data are shown as mean $\pm$ SEM. Statistical analysis was performed using two-way ANOVA followed by Tukey’s post hoc test in (B); *p 0.05, **p 0.01 and ***p 0.001. Fto, fat mass and obesity-associated protein; Nlrp3, nucleotide-binding oligomerization domain-, leucine-rich repeat-, and pyrin domain- containing receptor 3; MI, myocardial infarction; Anxa1, Annexin A1; IHC, immunohistochemistry; ASC, apoptosis-associated speck-like protein, cle-Caspase 1, cleaved-Caspase 1; N-Gsdmd, gasdermin D-N-terminal domain; Gsdmd, gasdermin D; SEM, standard error of the mean; ANOVA, analysis of variance.