Farrerol ameliorates diabetic cardiomyopathy by inhibiting ferroptosis via miR-29b-3p/SIRT1 signaling pathway in endothelial cells

Abstract

BACKGROUND

Diabetic cardiomyopathy (DCM) is the leading cause of cardiovascular disease-related mortality. Farrerol (FA) possesses anti-inflammatory and antioxidant properties. However, its role in regulating endothelial ferroptosis in DCM remains unknown.

AIM

To investigate the beneficial effects of FA on cardiac microvascular dysfunction in DCM from the perspective of ferroptosis in endothelial cells (ECs).

METHODS

The mice were fed a high-fat diet and injected with streptozotocin to induce DCM. DCM mice were orally administered FA (10 and 40 mg/kg/day) and a tail vein injection of the miR-29b-3p mimic or inhibitor for 24 weeks. Cardiac function and myocardial fibrosis were also analyzed. Cardiac microvascular function was assessed using immunofluorescence and transmission electron microscopy. Ferroptosis was analyzed using RNA sequencing, immunofluorescence, and western blotting.

RESULTS

FA administration improved cardiac function, alleviated myocardial fibrosis, strengthened endothelial barrier function, suppressed endothelial inflammation, and preserved the microvascular structure in DCM mice. This improvement was associated with the inhibition of endothelial ferroptosis and downregulation of miR-29b-3p in ECs. Similar efficacy was observed after tail vein injection of the miR-29b-3p inhibitor. Inhibition of miR-29b-3p in vivo showed an anti-cardiac fibrotic effect by improving microvascular dysfunction and ferroptosis in ECs, whereas overexpression of miR-29b-3p showed the opposite effects in DCM mice. Luciferase reporter assay revealed that miR-29b-3p binds to SIRT1. In cultured ECs, FA reduced high glucose and free fatty acid (HG/FFA)-induced lipid peroxidation and ferroptosis and inhibited endothelial-mediated inflammation. However, the overexpression of miR-29b-3p partially abolished the protective effects of FA against HG/FFA-induced injury in ECs. This finding suggests that the mechanism of action of FA in improving DCM is related to the downregulation of miR-29b-3p and activation of SIRT1 expression.

CONCLUSION

Therefore, FA has a potential therapeutic effect on cardiac microvascular dysfunction by suppressing EC ferroptosis through the miR-29b-3p/SIRT1 axis.

Key Words: Diabetic cardiomyopathy; Microvascular dysfunction; Farrerol; MiR-29b-3p; SIRT1

Core Tip: This study was the first to identify the miR-29b-3p/SIRT1 axis as a critical mechanism underlying the attenuation of endothelial ferroptosis and microvascular dysfunction of farrerol (FA) in diabetic cardiomyopathy (DCM). These results highlight the translational potential of targeting miR-29b-3p or administering FA for the treatment of DCM.

INTRODUCTION

Diabetes affects approximately 9.3% of the global population, with prevalence projected to increase to 10.9% by 2045[1]. In patients with diabetes, diabetic cardiomyopathy (DCM) has emerged as the predominant etiology of heart failure[2]. Notably, although preclinical investigations and clinical trials targeting DCM have markedly expanded in recent decades, the molecular pathogenesis underlying this disorder remains unclear. Cardiac microvascular endothelial cells (CMECs), pivotal functional units of the myocardial microvasculature, play an indispensable role in maintaining tissue perfusion homeostasis. Chronic exposure to hyperglycemia and dyslipidemia induces progressive endothelial injury and microcirculatory perturbations[3]. This mechanism precipitates cardiac functional deterioration and maladaptive myocardial remodeling in DCM. However, the precise molecular cascades driving CMEC dysfunction in the pathogenesis of DCM remain unclear.

Excessive lipid accumulation in the diabetic milieu induces the overgeneration of reactive oxygen species, creating a permissive environment for ferroptotic cell death, which exacerbates diabetic tissue damage[4]. Ferroptosis (distinctly iron-dependent and non-apoptotic) manifests as pathognomonic lipid peroxidation signatures[5] and is initiated without the involvement of death receptors, differentiating it from other regulated cell death processes[6]. Compelling evidence implicates ferroptosis in the pathogenesis of acute cardiovascular disorders, including myocardial infarction, ischemia/reperfusion injury, and chemotherapeutic cardiotoxicity[7-9]. Paradoxically, although these findings are predominantly derived from acute injury models, the spatiotemporal dynamics of ferroptosis in chronic metabolic disorders, particularly its contribution to DCM-associated cardiac dysfunction, remain unclear. This knowledge gap acquires clinical urgency given that diabetic iron overload exerts dual detrimental effects: Exacerbating insulin resistance through β-cell toxicity and amplifying cardiovascular damage via Fenton reaction-driven hydroxyl radical generation[10]. Endothelial ferroptosis is mechanistically associated with atherosclerotic plaque instability[11] and impaired angiogenesis in diabetic wounds[12]. Therefore, we hypothesized that ferroptosis is a pathognomonic feature of DCM progression, bridging metabolic derangements with microvascular decay and maladaptive remodeling.

Farrerol (FA), a bioactive flavanone isolated from Rhododendron dauricum L. (Ericaceae; traditional Chinese medicine “Man-shan-hong”), has been historically employed in Chinese pharmacopeia for respiratory disorders because of its mucoregulatory and antitussive properties[13]. Beyond its traditional applications, emerging evidence has revealed the pleiotropic therapeutic effects of FA spanning cardiovascular protection, neuromodulation, metabolic regulation, and oncological interventions. Mechanistically, FA exerts potent antioxidant and anti-inflammatory effects, as evidenced by the attenuation of neuroinflammation through microglial protection[14]. Recent studies have delineated its anti-ferroptotic capacity in collagenase-induced tendinopathy models[15] and its antifibrotic efficacy against cisplatin-induced nephropathy, positioning FA as a multitarget agent against cell death cascades[16]. A recent study found that DCM was alleviated by regulating the AMPK-mediated cardiac lipid metabolic pathways in rats with type 2 diabetes[17]. Moreover, FA plays a critical role in neuronal protection and inhibition of oxidative stress and ferroptosis via the Nrf2/Keap1 pathway[18]. Based on these premises, we hypothesized that endothelial ferroptosis is a critical driver of microvascular decay in DCM.

This study aimed to investigate the cytoprotective efficacy of FA against iron-dependent endothelial cell (EC) death and its underlying mechanisms. Using RNA sequencing, we found that treatment of DCM mice with FA upregulated SIRT1 and miR-29b-3p expression in the heart and that cell ferroptosis was involved in this regulation. SIRT1 is a nicotinamide adenine dinucleotide-dependent histone deacetylase that can modulate lipid peroxidation and ferroptosis[19,20]. It is regulated by microRNAs (miRNAs), which are short, noncoding RNAs that regulate the expression of target genes at the post-transcriptional level[21]. Thus, we hypothesized that FA ameliorates DCM by suppressing endothelial ferroptosis via the downregulation of miR-29b-3p and consequent activation of SIRT1 signaling.

MATERIALS AND METHODS

Animals

Male C57BL/6 mice (20 ± 3 g, 8 weeks old) were obtained from the Shanghai Slac Laboratory Animal Co., Ltd. (Shanghai, China). All procedures involving animals complied with the ARRIVE guidelines and the NIH Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of Shaoxing People’s Hospital (Approval No. 2022KY03901). All mice were housed under controlled conditions with a 14-hour light/8-hour dark cycle at a temperature of 24 °C and were provided unrestricted access to standard chow and water.

Animal experiments

Male C57BL/6 mice were randomly allocated to four groups (n = 8 each): Control, DCM, DCM + low FA, and DCM + high FA. The control group was fed a standard chow diet plus vehicle, whereas the DCM group was fed a high-fat diet (Hangs Biotechnology Co., Ltd.) for 24 weeks. Mice in the DCM group were intraperitoneally injected with streptozotocin (60 mg/kg, dissolved in 0.1 mmol/L citrate buffer, pH 4.4) for 5 consecutive days to induce DCM. The control animals received equivalent volumes of sterile citrate buffer. Fasting blood glucose was quantified weekly via tail vein sampling, and mice exhibiting sustained hyperglycemia (fasting blood glucose ≥ 16.7 mmol/L) were included in subsequent DCM analyses[22]. The diabetic mice were treated with different doses of FA. FA (#PHL82539; Sigma-Aldrich, United States) was dissolved in dimethyl sulfoxide (Sigma-Aldrich). Mice were orally treated with FA at a dose of 10 or 40 mg/kg/day in the DCM + low FA and DCM + high FA groups, respectively, based on previous studies[17,23].

To elucidate the effects of miR-29b-3p on DCM, we randomly divided 48 mice into six groups (n = 8 per group): NC, DCM, DCM + miRNA mimic negative control (DCM + mimic NC), DCM + inhibit negative control (DCM + inhibit NC), DCM + miR-29b-3p mimic, and DCM + miR-29b-3p inhibitor. The DCM + mimic NC, DCM + inhibit NC, DCM + miR-29b-3p mimic, and DCM + miR-29b-3p inhibitor groups were administered adeno-associated virus containing miR-29b-3p mimic, inhibitor, or corresponding oligonucleotides via tail vein injection once a week at a dose of 100 nmol in 100 μL of saline for 24 consecutive weeks. The dosage and administration were based on known pharmacokinetics/degradation rates of synthetic miRNAs in vivo[24-26] and according to the manufacturer’s instructions provided by Obio Technology Corp. Ltd. (Shanghai, China) and were used for all experiments. A flowchart of the experiment is shown in Supplementary Figure 1.

Echocardiography

For the echocardiography procedure, mice were anesthetized using 2%-4% sevoflurane and positioned supine on a heating pad. Transthoracic echocardiography was performed using the VisualSonics Vevo2100 system equipped with a 30 MHz central frequency probe to acquire M-mode and Doppler data. Cardiac function parameters such as left ventricular ejection fraction and left ventricular fractional shortening were subsequently calculated using VevoLAB software.

Histopathological analysis

Heart tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned to a thickness of 4 μm. The sections were subsequently deparaffinized, rehydrated, and stained with hematoxylin and eosin, Masson’s trichrome, and modified Sirius red (Solarbio, China) for collagen content assessment. Imaging was performed using a Leica microscope, and the images were analyzed using ImageJ software.

RNA sequencing and functional analysis

Total RNA was extracted from the cardiac tissue using TRIzol reagent, and its integrity was verified through agarose gel electrophoresis. Poly(A) RNA was fragmented and converted into cDNA using Super Script II. After pretreatment, the ligated products were amplified using polymerase chain reaction and subsequently subjected to 2 × 150 bp paired-end sequencing on an Illumina Novaseq 6000 platform. Bioinformatics analysis was performed using R (v4.3.0), with differential expression analyzed using DESeq2 (v1.38.1), applying a |log2Fold Change (log2FC)| > 1 and P < 0.05. Visualization of all differentially expressed genes (DEGs) was achieved through a Volcano plot, while a clustering heatmap of the top 20 upregulated genes was generated using the R packages “ggplot2” and “pheatmap”, respectively. Functional enrichment analysis, including Gene Ontology annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, was performed on the candidate genes using the “clusterProfiler” package[27].

Immunofluorescence staining

The hearts from various experimental groups were collected and perfused with 4% paraformaldehyde, followed by overnight fixation. The tissues were subsequently dehydrated in a graded sucrose series (10%, 20%, and 30%) and sectioned into slices of 5 μm thickness. For immunofluorescence analysis, sections were washed, permeabilized, and blocked. After blocking, the sections were incubated overnight with primary antibodies against CD31 (Catalog No. 11265), albumin (Catalog No. 16475), and Collagen I (Catalog No. 67288) from Proteintech, China. After three washes, secondary antibodies were applied. Sections were washed again and stained with 4′,6-diamidino-2-phenylindole for nuclear visualization. Images were captured using a confocal microscope (Stellaris 5; Leica).

Enzyme-linked immunosorbent assay

Homogenate supernatant of myocardial tissues (9 mL 0.9% physiological saline and 1 g tissue) was collected after centrifugation at 12000 rpm for 10 minutes, at 4 ℃. The contents of cardiac creatine kinase isozyme (CK-MB) and Lactate dehydrogenase (LDH) were tested in each group using the collected homogenate, and the experimental steps were performed strictly according to the kit instructions provided by Solarbio (China).

Detection of ferroptosis

Malondialdehyde (MDA; BC0025, Solarbio, China) levels were measured using a microplate reader (Molecular Devices, CA, United States). Intracellular ferrous iron (Fe²⁺) levels were determined according to the instructions provided with the iron assay kit (Cat.MAK025, Millipore Sigma).

Cell culture and treatment

Human umbilical vein ECs (HUVECs) were obtained from the National Collection of Authenticated Cell Cultures (Shanghai, China). These cells were cultured in an EC medium (ECM; ScienCell, United States) supplemented with 10% fetal bovine serum (Gibco). To simulate the effects of glucose and lipid exposure in vitro, we treated HUVECs with high glucose (HG, 25 mmol/L) and 0.5 mmol/L free fatty acids (FFA) for 72 hours, as previously described[28]. The FFA mixture consisted of bovine serum albumin (BSA)-conjugated palmitic acid and oleic acid at a 1:2 ratio. In the control group, HUVECs were cultured in ECM containing 5.5 mmol/L glucose, 1% BSA, and 19.5 mmol/L mannitol. To assess the protective effects of FA against HG/FFA-induced injury in HUVECs, we administered varying concentrations of FA (ranging from 1 to 100 μM) to the cells for 24 hours.

Cell transfection

MiR-29b-3p mimics (agonist), antagomiR-29b-3p (inhibitor), and scramble negative controls (miR-NC/inhibit-NC) were chemically synthesized by RiboBio Co., Ltd. (Guangzhou, China) with 2'-O-methyl modification for enhanced stability. HUVECs were plated in 6-well culture plates at a density of 1 × 10⁵ cells/well in complete ECM. At 70% confluency, cells were transfected with oligonucleotides (60 nM) or vector (0.5 μg) using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, United States). Post-transfection cultures were maintained for 48 h under normoglycemic conditions (5.5 mmol/L D-glucose) to modulate gene expression. Transfected HUVECs were subsequently exposed to HG/FFA. To knock down SIRT1, we transfected ECs with SIRT1-specific siRNA or scrambled control siRNA (GenePharma, Shanghai, China) using Lipofectamine 3000 according to the manufacturer’s protocol.

Cell proliferation assay

HUVECs transfected in a 96-well plate were stimulated with HG/FFA for 3 days, with or without FA treatment. After the addition of 10 μL of cell counting kit-8 (CCK-8; MedChemExpress, China) solution, the absorbance at 450 nm was measured using a microplate reader.

Endothelial monolayer permeability assay

To establish a confluent and tight EC monolayer, we seeded HUVECs at a density of 1 × 10⁴ cells into the upper compartment of a Transwell insert with a pore size of 0.4 μm (Corning Inc., United States) and maintained in culture for 3 days. Following this incubation period, 100 μL of fluorescein isothiocyanate-labeled BSA (FITC-BSA) at a concentration of 1 mg/mL (Solarbio, China) was added to the upper chamber, allowing it to permeate into the lower chamber via the paracellular pathway over 1 h. The fluorescence intensity of FITC-BSA in the upper and lower chambers was measured using a microplate reader. The albumin permeability coefficient was calculated as previously described[29].

Flow cytometric analysis of lipid peroxidation

BODIPY 581/591 C11 (Thermo Fisher Scientific) was used to assess cellular lipid peroxidation levels. In summary, treated HUVECs were subjected to trypsinization and subsequently resuspended in 400 μL ECM containing 2 μM C11-BODIPY (581/591) under dark conditions. The cells were collected for flow cytometry analysis using a CytoFLEX cytometer (Beckman Coulter) to measure the mean fluorescence intensity. The resulting data were analyzed using FlowJo software version 10.7.1.

Luciferase reporter assay

Luciferase assay was performed to confirm the interaction between miR-29b-3p and SIRT1. The eukaryotic miRNA plasmid and overexpression vector were constructed by GenePharma Co., Ltd. To generate SIRT1 wild-type (WT) and SIRT1 mutant (MUT) constructs, we inserted WT and MUT SIRT1 sequences containing miR-29b-3p binding sites into a pGL3 dual-luciferase vector. Subsequently, the miR-29b-3p mimic or negative control was co-transfected with SIRT1 WT or SIRT1 MUT. Luciferase assay was conducted 48 hours post-transfection. The fluorescence intensities of firefly and Renilla luciferases were measured using a multifunctional microplate reader.

RNA extraction and reverse transcription-quantitative PCR

Total RNA was extracted from cells and tissues using TRIzol reagent (Vazyme). miRNAs were reverse-transcribed into cDNA using the miRNA 1^st Strand cDNA Synthesis Kit (Vazyme) and specific miRNA reverse transcription primers. mRNAs were reverse-transcribed into cDNA using the HiScript II Reverse Transcription Kit (Vazyme) according to the manufacturer’s protocol. The relative mRNA and miRNA expression levels were normalized to those of GAPDH and U6, respectively. The primer sequences are listed in Supplementary Table 1.

Western blot

Proteins were prepared using RIPA lysis buffer (Beyotime, China) supplemented with protease inhibitors. Lysates were vortexed vigorously, incubated on ice for 30 minutes, and centrifuged at 12000 × g for 20 minutes at 4 °C to collect supernatants. The protein concentrations were determined using a BCA kit (Beyotime). Subsequently, 30 μg protein/sample was resolved on 10% Bis-Tris SDS-PAGE gels and transferred to PVDF membranes (Millipore). The membranes were blocked with 5% non-fat milk in TBST and incubated overnight with primary antibodies at 4 °C: Anti-SIRT1 (ab110304, Abcam), anti-intercellular adhesion molecule-1 (ICAM-1; ab282575, Abcam), anti- endothelial nitric oxide synthase (eNOS; ab199956, Abcam), anti-VE cadherin (ab33168, Abcam), anti-xCT (26864-1-AP, Proteintech), anti-GPX4 (67763-1-Ig, Proteintech), and anti-GAPDH (60004-1-Ig, Proteintech). On the second day, membranes were incubated for 1 hour with horseradish peroxidase-conjugated secondary antibodies for 1 hour at room temperature. The signals were detected using a SuperSignal™ West Pico ECL Substrate (Beyotime), and the quantification was performed using ImageJ software.

Statistical analysis

Statistical analyses were performed utilizing GraphPad Prism software 10.0, with data expressed as mean ± SD. All data were tested for normality (Shapiro-Wilk test) and homogeneity of variance (Levene’s test). Normally distributed data were analyzed using unpaired Student’s t-test (two groups) or one-way or two-way analysis of variance, followed by Tukey’s post-hoc test (multiple comparisons). Statistical significance was set at P < 0.05.

RESULTS

FA improved cardiac function in mice with DCM

In our study, DCM mice exhibited a decrease in left ventricular ejection fraction and left ventricular fractional shortening after long-term induction of diabetes. Meanwhile, oral administration of low and high FA significantly improved cardiac function in DCM mice (Figure 1A-C). Masson’s trichrome and Sirius red staining showed obvious perivascular fiber deposition in the DCM group, which decreased after FA treatment (Figure 1A, D and E). Mice treated with DCM showed increased serum levels of cardiac injury markers (CK-MB and LDH), whereas the addition of FA significantly reduced these markers (Figure 1F and G).

Figure 1 Treatment with Farrerol improved cardiac performance and histopathology in diabetic cardiomyopathy mice. A: Representative M-mode and pulse-wave Doppler echocardiograms of the mitral inflow (n = 8); B and C: Statistical analysis of the left ventricular ejection fraction and left ventricular fractional shortening in mice from indicated groups (n = 8); D and E: Interstitial fibrosis was detected by Masson's trichrome and Sirius red staining. Scale bars: 100 μm; F and G: Quantitative analysis of serum creatine kinase-MB and lactate dehydrogenase levels by ELISA (n = 8). ^aP < 0.05 vs NC; ^bP < 0.05 vs DCM. Eight biological replicates were included in the experiment, and the results are shown as scatter plots. FA: Farrerol; DCM: Diabetic cardiomyopathy.

FA alleviated cardiac microvascular disorder in mice with DCM

FA mitigated cardiac microvascular dysfunction in DCM mice. Building on the aforementioned findings and acknowledging the significance of cardiac microvascular dysfunction as a critical pathological component of DCM, we further evaluated the protective effects of FA on cardiac microvascular function. As illustrated in Figure 2A-C, cardiac microvascular vasodilation was impaired in DCM mice, as indicated by reduced microvascular density (Figure 2A and B). A compromised barrier function is another hallmark of cardiac microvascular injury. Albumin leakage was detected in the perivascular regions of diabetic hearts, a phenomenon ameliorated by FA treatment (Figure 2C). Furthermore, a decrease in eNOS expression and an increase in ICAM-1 expression were observed in diabetic hearts, suggesting impaired endothelium-dependent vasodilation. However, FA treatment enhanced cardiac microvascular perfusion by improving endothelium-dependent vasodilation (Figure 2D-F).

Figure 2 Treatment with Farrerol alleviated cardiac microvascular injury in diabetic cardiomyopathy mice. A and B: Cardiac microvascular density was determined using CD31-positive endothelial cells (ECs), and large vessels were excluded (n = 6 per fields); C: Perivascular leakage of albumin was detected by immunofluorescence staining and quantified. Bar = 50 μm; D-F: Protein expression of intercellular adhesion molecule-1 and endothelial nitric oxide synthase (eNOS) were detected by western blot and quantified (n = 3); G: Representative immunofluorescence images of endothelial anchoring junctions (VE-cadherin). Bar = 100 μm; H: In vitro endothelial monolayer permeability was determined based on FITC-BSA leakage (n = 6 per group); I and J: Representative images of immunofluorescence staining of Intercellular adhesion molecule-1 (VCAM-1) and quantitative analysis of VCAM-1 intensity. Bar = 20 μm. n = 6 per fields; K-M: Immunoblots and quantitative data of VCAM-1 and eNOS expression in high glucose and free fatty acid treated ECs (n = 3). ^aP < 0.05 vs NC; ^bP < 0.05 vs DCM or HG/FFA. FA: Farrerol; DCM: Diabetic cardiomyopathy; HG/FFA: High glucose and free fatty acid.

Subsequently, we determined whether FA improves endothelial function in vitro. FA concentrations exceeding 40 μM exhibit some cytotoxicity toward ECs; therefore, FA concentrations of 10 and 20 μM were used for subsequent studies (Supplementary Figure 2). As shown in Figure 2G and H, following HG/FFA injury, FA mitigated the leakage of FITC-BSA from the impaired endothelial monolayer by reinforcing the VE-cadherin junctions. In addition, FA significantly decreased the expression of VCAM-1 in ECs (Figure 2I and J). Accordingly, ICAM-1 expression was upregulated, and eNOS expression was reduced by HG/FFA treatment; however, this effect was significantly reversed by FA treatment (Figure 2K-M). In summary, these data indicated that FA exerts protective effects on cardiac microvascular function.

FA inhibits ferroptosis of ECs in DCM mice

To further investigate the underlying mechanisms by which FA ameliorates DCM, we performed bulk RNA sequencing of the hearts of FA-treated diabetic mice. Principal component analysis demonstrated high reproducibility of the samples within each experimental group, with the DCM + FA group exhibiting a distinct separation from the DCM group (Supplementary Figure 3A). RNA sequencing analysis of diabetic hearts identified 1498 upregulated and 534 downregulated DEGs, defined by a log2FC > 1 and a P value < 0.05. Comparatively, 451 upregulated and 1013 downregulated DEGs were identified between the groups (Supplementary Figure 3B and C). A heat map depicting 20 representative DEGs is presented in Figure 3A. Subsequently, the 998 DEGs were subjected to KEGG enrichment analysis. Comparative KEGG pathway analysis indicated significant enrichment of these DEGs in pathways related to coronavirus disease, cytokine-cytokine receptor interaction, the hypoxia-inducible factor signaling pathway, and ferroptosis (Figure 3B). Furthermore, gene set enrichment analysis confirmed the significant enrichment of the ferroptosis pathway (Supplementary Figure 3D). Several studies have identified ferroptosis as a novel therapeutic target for DCM[30,31]. Consistent with these findings, our data showed that the markers of ferroptosis (GPX4 and xCT) were reduced in DCM mice, whereas FA treatment significantly reversed these changes (Figure 3C-E). Additionally, the MDA content and Fe²⁺ levels were markedly elevated in diabetic hearts, whereas treatment with 10 and 40 mg/kg FA effectively reversed these changes (Figure 3F and G).

Figure 3 Farrerol treatment inhibited ferroptosis in diabetic hearts. A: RNA-sequencing was used to investigate the underlying mechanism of Farrerol on diabetic cardiomyopathy and a heatmap showed the top 20 differentially expressed genes (DEGs) between the groups; B: Kyoto Encyclopedia of Genes and Genome analysis of the top 100 DEGs between the groups; C-E: Western blot showed the changes of ferroptosis related proteins GPX4 and xCT in the four groups (n = 3); F: Relative expression of malondialdehyde in the hearts from indicated groups were determined by the commercial kit (n = 6); G: The content of Fe²⁺ in endothelial cells was assessed by the commercial kit (n = 6). ^aP < 0.05 vs NC; ^bP < 0.05 vs DCM. FA: Farrerol; DCM: Diabetic cardiomyopathy.

Furthermore, we established an in vitro model of DCM by inducing cellular injury using HG/FFAs based on previously published studies[32]. The results revealed that HG/FFAs significantly induced injury to cardiomyocytes, ECs, and fibroblasts in the three primary cardiac cell types. Notably, FA treatment markedly alleviated EC injury; however, it had limited protective effects against cardiomyocytes and fibroblasts (Figure 4A-C). Subsequently, we investigated HG/FFA-induced ferroptosis in ECs. FA treatment significantly reversed HG/FFA-induced downregulation of GPX4 and xCT in ECs (Figure 4D-F). Furthermore, the flow cytometry results demonstrated that HG/FFA treatment markedly increased lipid peroxidation in ECs, whereas FA intervention reduced lipid peroxidation in a concentration-dependent manner (Figure 4G). Concurrently, FA treatment significantly attenuated the HG/FFA-induced increase in MDA and iron ion levels (Figure 4H and I). In summary, FA may ameliorate DCM by suppressing EC ferroptosis.

Figure 4 Farrerol treatment inhibited ferroptosis in high glucose and free fatty acid treated endothelial cells. A-C: Cell viability was detected via cell counting kit-8 assay in cardiomyocytes, endothelial cells (ECs), and fibroblasts after indicated treatments (n = 6); D-F: Protein expression of GPX4 and xCT were detected in ECs upon high glucose and free fatty acid or different concentration of Farrerol treatment (n = 3); G: Flow cytometry is used to detect the changes in lipid peroxidation levels in ECs under different treatment conditions; H: Relative expression of malondialdehyde in the hearts from indicated groups were determined by the commercial kit (n = 6); I: The content of Fe²⁺ in ECs was assessed by the commercial kit (n = 6). ^aP < 0.05 vs NC; ^bP < 0.05 vs DCM. FA: Farrerol; DCM: Diabetic cardiomyopathy; HG/FFA: High glucose and free fatty acid.

Inhibition of miR-29b-3p suppress ferroptosis of ECs and cardiac fibrosis

To further elucidate the mechanism by which FA ameliorates DCM, we focused on the role of miR-29b-3p as the cardiac transcriptomic analyses revealed decreased expression of miR-29b-3p and SIRT1 in FA-treated hearts, with miRNAs known to play critical roles in DCM pathogenesis. In vitro experiments demonstrated that miR-29b-3p expression was downregulated in HG/FFA-treated cardiomyocytes and fibroblasts but upregulated in HG/FFA-stimulated ECs. FA intervention reversed the altered miR-29 expression in ECs but failed to restore miR-29b-3p levels in cardiomyocytes and fibroblasts (Figure 5A-C). Subsequently, we modulated miR-29b-3p expression in ECs by transfecting them with miR-29b-3p inhibitors or mimics to downregulate or upregulate miR-29b-3p, respectively (Figure 5D). Western blot analysis revealed that HG/FFA treatment reduced SIRT1 and the ferroptosis markers (GPX4 and xCT) in ECs. The upregulation of miR-29b-3p further suppressed, whereas its downregulation significantly increased the expression of SIRT1, GPX4, and xCT (Figure 5E-H). Additionally, overexpression of miR-29b-3p exacerbated the HG/FFA-induced elevation of MDA and iron levels, whereas inhibition of miR-29b-3p attenuated HG/FFA-driven ferroptosis of ECs (Figure 5I-J).

Figure 5 The effect of upregulation or downregulation of miR-29b-3p on high glucose and free fatty acid-induced ferroptosis of endothelial cells. A-C: Relative expression of miR-29b-3p in cardiomyocytes, endothelial cells (ECs), and fibroblasts was detected via reverse transcription-quantitative PCR analysis (n = 3); D: Relative expression of miR-29b-3p in ECs transfected with miR-29b-3p mimics, miR-29b-3p inhibitors and their corresponding negative controls (miR-NC and inhibit-NC, respectively); E-H: Protein expression of GPX4 and xCT were detected in ECs after indicated treatment (n = 3); I: Relative expression of malondialdehyde in the hearts from indicated groups were determined by the commercial kit (n = 6); J: The content of Fe²⁺ in ECs was assessed by the commercial kit (n = 6). ^aP < 0.05 vs NC; ^bP < 0.05 vs HG/FFA. FA: Farrerol; DCM: Diabetic cardiomyopathy; HG/FFA: High glucose and free fatty acid.

Subsequently, we validated the protective effects of miR-29b-3p on DCM using in vivo experiments. Echocardiography, hematoxylin and eosin staining, Masson’s trichrome staining, and Sirius red staining demonstrated that mice transfected with mimic NC (negative control mimic) and NC (negative control inhibitor) showed no statistically significant differences in cardiac function or collagen deposition levels compared with those in the DCM group. However, DCM mice transfected with the miR-29b-3p mimic exhibited worse cardiac function and increased collagen deposition than the DCM group, whereas DCM mice transfected with the miR-29b-3p inhibitor displayed improved cardiac function and reduced collagen deposition (Figure 6A-E). Similarly, serum BNP levels were elevated in the miR-29b-3p mimic group but decreased in the miR-29b-3p inhibitor group compared with those in the DCM group (Figure 6F and G). Western blot analysis of cardiac tissues revealed that miR-29b-3p overexpression further reduced SIRT1 expression and exacerbated the decline in ferroptosis-related proteins (GPX4 and xCT). Conversely, inhibition of miR-29b-3p upregulated the expression of SIRT1 and ferroptosis-associated proteins (Figure 6H-K).

Figure 6 Effects of upregulation or downregulation of miR-29b-3p on histopathology and ferroptosis in diabetic cardiomyopathy mice. A: Representative echocardiographic images for six groups of mice; B and C: Quantification of Left ventricular ejection fraction and left ventricular fractional shortening in mice from specified groups (n = 8); D and E: Masson's trichrome and Sirius red staining was used to detect cardiac fibrosis. Scale bars: 100 μm; F and G: Analysis of serum myocardial enzyme activity creatine kinase-MB and lactate dehydrogenase levels (n = 8); H-K: Representative blot images and quantitative analysis of SIRT1, GPX4 and xCT expression in diabetic hearts from the indicated groups (n = 3). ^aP < 0.05 vs NC; ^bP < 0.05 vs DCM. FA: Farrerol; DCM: Diabetic cardiomyopathy; LVEF: Left ventricular ejection fraction; LVFS: Left ventricular fractional shortening; CK-MB: Creatine kinase-MB; LDH: Lactate dehydrogenase.

Subsequently, we investigated the role of miR-29b-3p in microvascular disorders. We found that overexpression of miR-29b-3p further reduced microvascular density and increased albumin leakage in diabetic hearts, whereas inhibition of miR-29b-3p increased microvascular density and reduced albumin leakage (Figure 7A-C). Accordingly, the overexpression of miR-29b-3p further reduced the protein expression of eNOS and VE-cadherin and increased ICAM-1 protein expression in diabetic hearts. In contrast, the inhibition of miR-29b-3p increased the expression of eNOS and ICAM-1 and reduced the expression of ICAM-1 in diabetic hearts (Figure 7D-G). These results indicate that overexpression of miR-29b-3p aggravates DCM by promoting ferroptosis, whereas inhibition of miR-29b-3p improves DCM by inhibiting ferroptosis.

Figure 7 The effects of upregulation or downregulation of miR-29b-3p on cardiac microvascular injury in diabetic cardiomyopathy mice. A and B: Cardiac microvascular density was determined using CD31-positive endothelial cells, and large vessels were excluded (n = 6); C: Perivascular leakage of albumin was detected by immunofluorescence staining and quantified. Bar = 50 μm; D-G: Protein expression of intercellular adhesion molecule-1, VE-cadherin and endothelial nitric oxide synthase were detected by western blot and quantified (n = 3). ^aP < 0.05 vs NC; ^bP < 0.05 vs DCM. DCM: Diabetic cardiomyopathy; ECs: Endothelial cells; ICAM-1: Intercellular adhesion molecule-1; eNOS: Endothelial nitric oxide synthase.

FA improves ferroptosis and dysfunction of ECs by regulating miR-29b-3p

To investigate the role of miR-29b-3p in mediating the protective effects of FA against ferroptosis and EC dysfunction, we overexpressed miR-29b-3p in FA-treated ECs. Our findings revealed that miR-29b-3p overexpression markedly amplified the HG/FFA-induced suppression of SIRT1 and the ferroptosis-related markers xCT and GPX4. Using bioinformatic analysis, we hypothesized that SIRT1 and miR-29b-3p would interact (Supplementary Figure 4A). Dual-luciferase reporter analysis demonstrated that the relative luciferase activity was drastically reduced in ECs co-transfected with SIRT1-WT and miR-29b-3p mimics (Supplementary Figure 4B). miR-29b-3p overexpressed ECs showed lower expression of SIRT1 than the control group; inhibition of miR-29b-3p significantly upregulated SIRT1 expression (Supplementary Figure 4C). MiR-29b-3p overexpression partially negated the ability of FA to restore SIRT1, xCT, and GPX4 expression in ECs (Figure 8A-D). Immunofluorescence staining further demonstrated that SIRT1 was localized to the cytoplasmic and nuclear compartments of ECs. Notably, miR-29b-3p overexpression exacerbated the HG/FFA-driven reduction in SIRT1 fluorescence intensity and diminished the ability of FA to upregulate SIRT1 expression (Figure 8E and F). Flow cytometry analysis of lipid peroxidation levels showed that miR-29b-3p overexpression intensified HG/FFA-induced lipid peroxidation and partially reversed the suppression of this oxidative process by FA (Figure 8G and H). We assessed the functional effects of miR-29b-3p on ECs under HG/FFA stress. miR-29b-3p overexpression aggravated the HG/FFA-induced downregulation of VE-cadherin and partially abolished the restorative effect of FA on VE-cadherin expression and leakage of FITC-BSA (Figure 9A and B). Similarly, miR-29b-3p overexpression reduced VCAM-1 expression in HG/FFA-stimulated ECs and reduced the ability of FA to increase VCAM-1 Levels (Figure 9C and D). Additionally, transfection with the miR-29b-3p mimic lowered ICAM-1 expression, enhanced eNOS levels in HG/FFA-treated ECs, and counteracted the regulatory effects of FA on these markers (Figure 9E-G).

Figure 8 Farrerol inhibits high glucose and free fatty acid-induced ferroptosis via inhibition of miR-29b-3p in endothelial cells. A-D: Protein expression of SIRT1, GPX4 and xCT were detected in endothelial cells (ECs) upon high glucose and free fatty acid, together with Farrerol or miR-29b-3p (n = 3); E and F: Representative immunofluorescence images of SIRT1 in ECs from indicated group. Bar = 50 μm. n = 6; G-H: Flow cytometry is used to detect the changes in lipid peroxidation levels in ECs under different treatment conditions (n = 6). ^aP < 0.05 vs NC; ^bP < 0.05 vs HG/FFA + mimic NC; ^cP < 0.05 vs HG/FFA + miR-29b-3p mimic; ^dP < 0.05 vs HG/FFA + FA + mimic NC. HG/FFA: High glucose and free fatty acid; FA: Farrerol.

Figure 9 Farrerol alleviates high glucose and free fatty acid-induced endothelial cell dysfunction via inhibition of miR-29b-3p. A and B: Representative immunofluorescence images and quantitative analysis of the intensity endothelial anchoring junctions (VE-cadherin). Bar = 100 μm. n = 6; C and D: Representative images of immunofluorescence staining of intercellular adhesion molecule-1 (VCAM-1) and quantitative analysis of VCAM-1 intensity. Bar = 20 μm. n = 6; E-G: Western blots analysis and quantitative data of VCAM-1 and endothelial nitric oxide synthase expression in high glucose and free fatty acid treated endothelial cells (n = 3). ^aP < 0.05 vs NC; ^bP < 0.05 vs HG/FFA + mimic NC; ^cP < 0.05 vs HG/FFA + miR-29b-3p mimic; ^dP < 0.05 vs HG/FFA + FA + mimic NC. eNOS: Endothelial nitric oxide synthase; ICAM-1: Intercellular adhesion molecule-1; HG/FFA: High glucose and free fatty acid; FA: Farrerol.

FA improves ferroptosis and dysfunction of ECs by regulating the miR-29b-3p/SIRT1 axis

To further define the role of the miR-29b-3p/SIRT1 axis in the protective effects of FA against HG/FFA-induced EC dysfunction, we manipulated the expression of SIRT1 in ECs. As shown in Figure 10A, SIRT1 knockdown was achieved by transfecting cells with specific siRNAs, and reduced SIRT1 protein levels were confirmed by western blotting. The CCK-8 assay results revealed that SIRT1 silencing markedly abolished the protective effect of FA on cell viability in HG/FFA-treated ECs (Figure 10B). Furthermore, SIRT1 knockdown abrogated the inhibitory effects of FA on EC ferroptosis (Figure 10C-E). Figure 10F-G shows that SIRT1 silencing partially reversed the protection of VE-cadherin junctions and reduced the mitigation of FITC-BSA leakage from injured ECs by FA. To confirm that the effect of miR-29b-3p on ECs depends on SIRT1, we knocked down SIRT1 (Figure 10H). The inhibition of miR-29b-3p alleviated HG/FFA-induced EC injury; however, this protective effect was reversed by SIRT1 silencing. Specifically, SIRT1 knockdown abolished the suppression of HG/FFA-induced decreases in the ferroptosis markers GPX4 and xCT by the miR-29b-3p inhibitor (Figure 10I-K). Additionally, the miR-29b-3p inhibitor-mediated alleviation of VE-cadherin junction disruption and FITC-BSA leakage in injured ECs was reversed by SIRT1 silencing (Figure 10L and M). Collectively, these results demonstrate that FA mitigates HG/FFA-induced EC injury through the miR-29b-3p/SIRT1 signaling axis.

Figure 10  SIRT1 was involved in the protective effects of Farrerol and miR-29b-3p on high glucose and free fatty acid-induced endothelial cells dysfunction. A: Knockdown of SIRT1 was achieved by transfected siRNA into Human umbilical vein Endothelial cells and confirmed by western blot analysis (n = 3); B: Cell viability was detected via cell counting kit-8 (CCK-8) assay in endothelial cells (ECs) after indicated treatments (n = 6); C-E: Western blots analysis and quantitative data of GPX4 and xCT expression in high glucose and free fatty acid (HG/FFA) treated ECs (n = 3); F: Representative immunofluorescence images of the intensity endothelial anchoring junctions (VE-cadherin). Bar = 100 μm; G: In vitro endothelial monolayer permeability was determined based on FITC-BSA leakage (n = 6 per group); H: Cell viability was detected via CCK-8 assay in ECs with SIRT silencing (n = 6); I-K: Western blots analysis and quantitative data of GPX4 and xCT expression in HG/FFA treated ECs (n = 3); L: Representative immunofluorescence images of the intensity endothelial anchoring junctions (VE-cadherin). Bar = 100 μm; M: In vitro endothelial monolayer permeability was determined based on FITC-BSA leakage (n = 6 per group). ^aP < 0.05 vs NC; ^bP < 0.05 vs HG/FFA; ^cP < 0.05 vs HG/FFA + miR-29b-3p mimic. HG/FFA: High glucose and free fatty acid.

DISCUSSION

Here, we demonstrated that FA ameliorated microcirculatory dysfunction in the hearts of mice with DCM by inhibiting ferroptosis in ECs, likely by regulating the miR-29b-3p/SIRT1 axis. These findings provide novel insights into the mechanisms underlying microcirculatory dysfunction in DCM and offer experimental evidence for the therapeutic potential of FA in treating DCM.

FA is a plant polyphenol compound found abundantly in certain fruits, vegetables, and herbal plants and possesses therapeutic effects across diverse disease spectra[13,33,34]. These therapeutic effects are primarily due to their free radical neutralization, antioxidant, and anti-inflammatory properties. Moreover, FA alleviates cerebral ischemia/reperfusion injury in vivo[35] and protects against cardiac hypertrophy and remodeling in angiotensin II-induced hearts[23]. Previous studies confirmed the protective effect of FA on the cardio-cerebrovascular system. Notably, FA ameliorates the pathogenesis of metabolically associated fatty liver disease by alleviating insulin resistance and suppressing hepatic lipid accumulation[36]. Furthermore, FA alleviates collagenase-induced tendinopathy by inhibiting ferroptosis[15]. These findings demonstrate the protective effects of FA against metabolic diseases. Our study showed that FA treatment improved cardiac microcirculation dysfunction in DCM mice in a concentration-dependent manner. While previous studies have shown that FA can reduce inflammation and apoptosis and suppress lipid accumulation in cardiomyocytes[17,23], we focused on the effect of FA on EC injury in DCM and its influence on cardiac microcirculation. To our knowledge, this is the first study to investigate its role in specifically protecting the cardiac microvasculature by targeting endothelial ferroptosis. Microvascular dysfunction is a critical early event in DCM development.

Studies have found that even in the absence of significant coronary artery disease, chest pain and ischemia are common in patients with diabetes. This condition, known as nonobstructive coronary artery disease, is caused by microvascular dysfunction due to coronary artery spasms[37,38]. Cardiac microvascular ECs are integral structural and functional units within the cardiac tissue and play a pivotal role in maintaining microvascular perfusion homeostasis[39]. Chronic exposure to hyperglycemia and dysregulated lipid metabolism induces persistent endothelial injury and microcirculatory dysfunction, which ultimately contribute to the progression of cardiac dysfunction and pathological myocardial remodeling[40]. In our experiments, DCM mice exhibited reduced cardiac function and myocardial collagen deposition, indicating cardiac remodeling. In addition, diabetic hearts exhibit increased lipid peroxidation and ferroptosis in ECs. Several studies have shown that inhibiting ferroptosis in ECs plays a crucial role in alleviating cardiac microvascular dysfunction associated with diabetes[41-43]. The addition of FA alleviated endothelial and microvascular injury and inhibited EC ferroptosis in vivo and in vitro.

miRNAs, which are small, noncoding, single-stranded RNA molecules that function as essential regulators of key biological processes, such as metabolic activity, cell proliferation, and programmed cell death, thereby maintaining systemic homeostasis[44]. Multiple cardiac-specific miRNAs are significantly dysregulated in diabetes, and emerging evidence has characterized the mechanistic involvement of distinct miRNAs in driving DCM progression by modulating pathological pathways[45]. Recently, miR29b3p was found to inhibit the activation of hepatic stellate cells and migrate liver fibrosis[46]. In addition, inhibition of miR29b3p was found to protect human trabecular meshwork cells against oxidative injury[47]. These results suggest that miR-29b-3p plays an important role in oxidative stress-mediated ferroptosis and fibrosis-related diseases. However, the expression pattern and biological function of miR-29b-3p in DCM remain unclear. Here, for the first time, we found that the expression of miR-29b-3p differed among different cell types in diabetic hearts. Inhibition of miR-29b-3p alleviates cardiac fibrosis in DCM mice, improves cardiac microcirculatory dysfunction, and suppresses ferroptosis in ECs. To further investigate the mechanisms by which miR-29b-3p exerts its antifibrotic effects in DCM, we focused on SIRT1, which protects EC homeostasis by activating antioxidant pathways, inhibiting inflammation, and maintaining mitochondrial function[48]. In the present study, we identified SIRT1 as a direct target of miR-29b-3p in ECs. miR-29b-3p protected ECs from HG/FA-induced dysfunction by targeting SIRT1. A recent study reported that morin alleviates DCM by suppressing ferroptosis via the SIRT1/p53/SLC7A11 signaling pathway[49]. Similarly, irisin increases SIRT1 to upregulate SLC7A11 and GPX4 expression, thereby inhibiting ferroptosis and protecting cardiomyocytes against injury caused by HG[50]. These data indicate that SIRT1 is an effective target for ferroptosis in DCM. More importantly, Guo et al[51] mechanistically delineated that miR-34a ablation attenuates myocardial fibrosis through SIRT1/PGC-1α/FNDC5 signaling axis activation. In this study, we demonstrated that SIRT1 knockdown abolished the suppressive effects of FA on HG/FFA-induced ferroptosis markers and inflammatory cytokines, thereby confirming the necessity of SIRT1 for the beneficial effects of FA. Additionally, SIRT1 knockdown partially reversed the protective effects of miR-29b-3p on the function of ECs. Taken together, these results suggest that miR-29b-3p is an upstream regulator of SIRT1 and that FA is a natural antagonist of miR29b3p that effectively prevents ferroptosis in ECs.

The present study had certain limitations. First, FA possesses pharmacological activities, including anti-inflammatory and antioxidative effects[13,15,35]. The mechanism by which FA regulates microvascular injury in DCM is complex and requires further investigation of other signal transduction networks. Second, the expression of miR-29b-3p varies across different cell types. While our data support a role for endothelial miR-29b-3p/SIRT1 signaling in the protective mechanism of FA, systemic delivery of miR-29b-3p modulators may affect non-target cell types. Future studies should employ endothelial-specific miR-29b-3p knockout models to verify cell-autonomous effects. Further investigation is warranted to delineate cell type-dependent regulatory mechanisms and develop endothelial-targeted nanoparticle systems for delivering miR-29b-3p modulators. Third, although we assessed key ferroptosis markers (GPX4, cXT, and lipid peroxidation), a limitation of our study was the absence of data for other specific markers, such as acyl-coenzyme A synthetase long-chain family member 4 and 4-hydroxynonenal, and the lack of rescue experiments using ferroptosis inhibitors. Future studies incorporating these approaches should confirm the specific role of ferroptosis inhibition in the mechanism of FA. Finally, we used a high-fat diet and streptozotocin to induce DCM rather than db/db mice because the latter had a higher weight and required large amounts of FA and miR-29b-3p. Finally, the injection of miR-29b-3p was not limited to ECs but also to other cells. These issues should be addressed in future studies using cardiac-specific deletion or overexpression of miR-29b-3p in mouse lines.

CONCLUSION

Collectively, this study demonstrates that FA provides cardio-protection against cardiac microvascular dysfunction by inhibiting EC ferroptosis via the miR-29b-3p/SIRT1 pathway. These findings suggest a pivotal role of miR-29b-3p-mediated ferroptosis of ECs in diabetic hearts, providing a new potential therapeutic target for DCM.