Circular RNA circFTO promotes pressure overload-induced cardiac hypertrophy by encoding a novel protein FTO-36aa

Abstract

Background

Pathological cardiac hypertrophy leading to ventricular remodeling poses a significant threat to human health. Circular RNAs (circRNAs) play a potential role in the dysregulation of cardiac hypertrophy, and recent evidence highlights their translational ability in various diseases. However, it is not clear whether circRNAs play a protein-coding role in myocardial hypertrophy and ventricular remodeling. This study aimed to investigate the role of circRNA derived from the fat mass and obesity-associated (FTO) gene (circFTO), a translatable circRNA, and the circFTO-encoded a 36 amino acid protein (FTO-36aa) in the pathogenesis of myocardial hypertrophy.

Methods

A transverse aortic constriction (TAC)-induced hypertrophy mouse model was established. The heart function of the C57BL/6 mice was evaluated. Myocardial structure injury and fibrosis were analyzed by hematoxylin and eosin (H&E) staining and Masson staining. CircRNA microarray assays were used to screen the dysregulated circRNAs. The recombinant adenovirus-associated virus (AAV) was constructed to overexpress or knockdown FTO protein or circFTO. Mass spectrometry analyses, dual-luciferase reporter assays, and polysome profiling analyses were performed to detect the FTO-36aa.

Results

The study identified dysregulated circRNAs in sham and TAC models, and found that an upregulated circRNA, circFTO, is generated from the back-splicing of FTO exon 5 and exon 7. The silencing of circFTO by AAV significantly weakened the TAC-induced hypertrophy phenotype. The study also identified a novel protein, FTO-36aa, coded by circFTO, that caused the pro-hypertrophy effect of circFTO. FTO-36aa promoted the ubiquitination-mediated protein degradation of FTO, which suppressed the demethylation of RNA, elevating the global N6-methyladenosine (m⁶A) methylation. Further, the m⁶A reader, IGF2BP2, recognized the circFTO/FTO-36aa elevated m⁶A methylation and increased the messenger RNA (mRNA) stabilities of the m⁶A methylated hypertrophic genes.

Conclusions

Overall, this study shed light on the functional importance of alternative splicing-generated circFTO and its coded FTO-36aa during myocardial hypertrophy. The findings provide fundamental insights into the mechanisms of m⁶A methylation regulation in hypertrophic cardiomyocytes.

Highlight box.

Key findings

• This study found that the circular RNA (circRNA) derived from the fat mass and obesity-associated (FTO) gene (circFTO) is significantly upregulated in pressure overload-induced cardiac hypertrophy. CircFTO-encoded a 36 amino acid protein (FTO-36aa) promotes the ubiquitination-mediated degradation of the FTO protein. Reduced FTO levels diminish the demethylase activity, leading to elevated global N6-methyladenosine (m⁶A) RNA methylation. The m⁶A reader IGF2BP2 recognizes these methylated RNAs and stabilizes hypertrophic gene messenger RNAs (mRNAs) and exacerbating cardiac hypertrophy. Silencing circFTO or blocking FTO-36aa attenuates hypertrophy in mice, improving cardiac function and reducing fibrosis.

What is known, and what is new?

• CircRNAs are implicated in myocardial hypertrophy. FTO, a m6A demethylase, regulates cardiac remodeling, and its downregulation is correlated with increased m⁶A methylation in hypertrophy.

• This study showed that circFTO, a translatable circRNA, encodes FTO-36aa, a protein that post-translationally suppresses FTO protein, linking circRNA translation to m6A-mediated hypertrophy. It identified a circRNA-protein-m⁶A axis whereby IGF2BP2 stabilizes methylated hypertrophic mRNAs, providing a new mechanistic layer in disease pathogenesis.

What is the implication, and what should change now?

• CircRNAs may exert pathological effects via encoded proteins. The therapeutic targeting of circFTO/FTO-36aa or IGF2BP2 could mitigate m⁶A-driven hypertrophy. Clinicians should consider circRNA-protein interactions as potential therapeutic targets. Research should be conducted to investigate the upstream regulators of circFTO biogenesis and explore cross-talk between m⁶A and other epigenetic modifications in hypertrophy. Before these findings can be applied to anti-hypertrophic strategies, the circFTO-FTO-IGF2BP2 axis must be validated in human models and larger animal studies to assess its safety and efficacy.

Introduction

Pathological myocardial hypertrophy is the pathophysiological basis of ventricular remodeling, that causes insufficient cardiac output and chronic heart failure (1,2). Therefore, preventing myocardial hypertrophy is crucial in preventing the development of ventricular remodeling and enhancing the prognosis of patients with heart failure (3-6).

Circular RNAs (circRNAs) are single-stranded endogenous non-coding RNAs with no free ends that create a unique ring structure through covalent closure (7). They are characterized by stability, abundant expression, high conservatism, cell specificity, tissue specificity (8), developmental specificity, etc. (9-12). Due to their tissue-specific expression and fluctuating characteristics during disease, circRNAs are involved in the pathophysiological mechanism of myocardial hypertrophy and ventricular remodeling (13,14). CircRNA_010567 prevents interstitial fibrosis after myocardial infarction by inhibiting miR-141 expression (15). Researchers on myocardial hypertrophy induced by pressure overload via transverse aortic constriction (TAC) have established a mouse model of pressure overload myocardial hypertrophy (16). Heart-related circRNA can sponge miR-223, enhance the expression of apoptotic inhibitor protein, and hamper TAC-induced cardiac hypertrophy (17). Another study has reported that circSlc8a1 is highly expressed in mouse and human heart tissues (18). CircRNA IGF1R promotes cardiac repair via activating beta-catenin signaling by interacting with DDX5 in mice after ischemic insults (19). Animal experiments have confirmed that the knockdown of circSlc8a1 can delay the myocardial hypertrophy caused by pressure load in mice by targeting miR-133a, and circSlc8a1overexpression ultimately leads to heart failure (18). CircMIRIAF aggravates myocardial ischemia-reperfusion injury by targeting the miR-544/WDR12 axis (20). Previous circRNA ArrayStar microarray research has revealed that the circRNA expression profile changed significantly during myocardial hypertrophy, such that 235 circRNAs were significantly upregulated and 84 circRNAs were significantly downregulated during myocardial hypertrophy. However, research on procedures for reconstructing circRNA with significant regulatory effects from numerous circRNAs with different levels of expression in the ventricle is limited.

Most circRNAs are expressed by familiar protein-coding genes. CircRNAs are composed of one or more exons and/or introns, and the composition of circRNAs includes exon circRNAs, intron circRNAs, and exon-intron circRNAs (21). Studies on circRNAs are limited. CircRNAs function to sponge microRNAs (miRNAs) (20), interact with RNA-binding proteins (19), regulate protein levels before and after transcription (22), and a few circRNAs could code proteins (23-26). Chen and Sarnow found that the circRNA inserted into the internal ribosome entry site (IRES) showed translation ability, proving that not all circRNAs are non-coding RNAs (27).

Extensive research on circRNA has led to breakthroughs (28). Endogenous circRNA encoding uses variable splicing, and an open reading frame (ORF) is a template that manages 5'-cap-independent protein translation via an IRES-mediated, overlapping genetic code (29,30). Covalently bonded circRNAs are more stable than linear RNA, and are less susceptible to exonuclease hydrolysis, making the median half-life of circRNAs in mammalian cells at least 2.5 times longer than that of linear messenger RNA (mRNA) isomers (31-34). However, research on the protein-coding role of circRNA in myocardial hypertrophy and ventricular remodeling is rare. This study investigated the potential translation ability of the top upregulated circRNAs in hypertrophic hearts, and their role in myocardial hypertrophy and ventricular remodeling.

Currently, epigenomics, especially the N6-methyladenosine (m⁶A) modification, is a popular area of research (35). The m⁶A modification is a common internal modification of the mRNA termination codon and 5'-untranslated region, which regulating RNA degradation, shearing, localization, transportation, and translation at the post-transcriptional level (36). It regulates abnormal embryonic development, tumorigenesis, immune diseases, blood diseases, neurological diseases, obesity, osteoporosis, etc. (37-39). The first identified m⁶A demethylase, fat mass and obesity-associated (FTO) protein, revealed the dynamic reversible regulation of m⁶A (40). By removing m⁶A methyl, FTO protein assists in the sepsis process of DNA damage repair and pre-mRNA splicing, stimulates tumor proliferation, accelerates myocyte differentiation, and inhibits hippocampal memory formation (41,42). Mathiyalagan et al. reported that the m⁶A methylation of mRNA was significantly higher in ischemic myocardium than non-ischemic myocardium (43). The expression of m⁶A demethylase (FTO) was reduced in ischemic offspring and mice (43). Under hypoxia, the expression of FTO in cardiomyocytes is downregulated, increasing the mRNA methylation, decreasing the expression of hypertrophic genes in the myocardium, and impairing cardiomyocyte function (43). However, the upstream regulatory mechanism of the dysfunctional FTO and m⁶A in myocardial hypertrophy remains unclear.

The aim of this study was to investigate whether dysregulated circRNAs play a role in suppressing m⁶A modification during myocardial hypertrophy. RNA-sequencing screening of the irregulated circRNAs in sham and TAC mice was conducted, and the results showed that mmu_circ_0001717 [circRNA derived from the FTO gene (circFTO)], which was generated from the host FTO gene between exon 5 and exon 7, was increased in the TAC mice. Additionally, a novel protein [circFTO-encoded a 36 amino acid protein (FTO-36aa)] was identified. This protein promoted ubiquitination-mediated protein degradation of the FTO protein. The results suggest that targeting circFTO-suppressed FTO signaling could be a potential therapeutic strategy for treating pathological myocardial hypertrophy. We present this article in accordance with the ARRIVE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1462/rc).

Methods

Animals

A total of 72 adult male C57BL/6 mice (8–10 weeks old) were purchased from the Shanghai Laboratory Animal Center at the Chinese Academy of Sciences (Shanghai, China). Block randomization was used to allocate the mice to control and treatment groups, ensuring equal group sizes. Specifically, six mice were allocated to each experimental group, with four groups per experiment (the control + three treatment groups), totaling 24 mice per experiment (six mice/group × four groups). Three independent experiments were performed in this study; thus, the study used 72 adult male C57BL/6 mice in total. The investigator responsible for randomization was aware of the group assignments.

The mice were housed in a pathogen-free environment. All the animal procedures followed National Institutes of Health (NIH) guidelines for the care and use of animals and were approved by the Animal Care and Use Committee of The First Affiliated Hospital, Nanchang University (No. 2020-62). A protocol was prepared before the study without registration. For the invasive surgeries, the mice were anesthetized using isoflurane (induction: 3–4%, maintenance: 1–2%) to ensure unconsciousness during the procedure. Post-operative analgesia (buprenorphine, 0.1 mg/kg) was administered subcutaneously every 8–12 h for 48 h. Humane endpoints were strictly followed. Euthanasia was performed via carbon dioxide (CO₂) inhalation followed by cervical dislocation to confirm death. No severe adverse events were observed during the study.

TAC

A TAC-induced hypertrophy model was established (44). The C57BL/6 mice were anesthetized by ketamine (100 mg/kg) intraperitoneal injection and xylazine (10 mg/kg), and aseptic surgery was performed. The mouse was laid on a magnetic stainless steel surgical board. The upper half of the sternum was divided using scissors in the middle to remove the thymus, and the aortic arch was cut after the skin incision. A 7-0 silk suture was used to tightly bind the aortic arch around a blunt needle (27 gauge), and was then quickly removed. The sternotomy and the skin incision were sealed with 5-0 stitches. Except for the aorta ligation, the sham-operated animals underwent a similar procedure. The mice were kept warm on a heating pad throughout the procedure and recovery.

The heart contractile function was assessed by echocardiography 4 weeks after surgery. Prior to any procedures, the mice were anesthetized to minimize pain and distress. Isoflurane (2–4% in oxygen) was administered via a precision vaporizer in an induction chamber until the mice reached a surgical plane of anesthesia. The depth of anesthesia was monitored by observing the absence of the toe-pinch reflex. To minimize suffering, euthanasia was performed under deep anesthesia. The chosen method for sacrifice was cervical dislocation, which is recommended as a quick and humane method for small rodents such as mice. The neck was gently grasped, and a swift and firm pull was applied to dislocate the cervical vertebrae.

Histology

The histology analysis of the isolated hearts was performed as described previously (44). The isolated hearts were fixed in 10% formalin, dehydrated, and paraffin-embedded. The 5-µm sections were stained with hematoxylin and eosin (H&E), and wheat germ agglutinin (WGA; Alexa Fluor 488 conjugate; Thermo Fisher, Rockford, IL, USA). Masson (HT15-1KT; Sigma-Aldrich, St. Louis, MO, USA) staining was performed as per the instruction manual to determine collagen deposition. Image Pro-Plus version 6.0 (Media Cybernetics, Rockville, MD, USA) image analysis software was used to examine the slides from a microscopic perspective. The fibrotic area was determined using the myocardial collagen area and the field area.

Echocardiographic measurement

Echocardiographic measurement of the mice was conducted (44). The C57BL/6 mice were anesthetized briefly by a pentobarbital sodium (40 mg/kg) injection, and fixed in a prone position on a heating pad. The heart rate was monitored with a standard limb lead-II electrocardiogram and maintained at 50±5 per min during echocardiography. Cardiac functions were evaluated by transthoracic echocardiography using an ultrasound machine (Visual Sonics Inc., Toronto, ON, Canada) with a 716 probe. A high-resolution electrocardiograph system was used to calculate the left ventricular end-systolic dimension (LVESD), left ventricular end-diastolic volume (LVEDV), left ventricular ejection fraction (LVEF), and left ventricular fractional shortening (LVFS). The results were averaged over three consecutive cardiac cycles.

CircRNA microarray assays

TRIzol reagent (Invitrogen, Carlsbad, CA, USA) was used to extract total RNA from the heart tissues of the mice, which then underwent a quantified analysis (NanoDrop, Thermo Fisher, Waltham, MA, USA). The microarray hybridization was conducted using the standard protocol of ArrayStar (Rockville, MD, USA). Fluorescent complementary RNAs were synthesized from the purified RNAs and hybridized in a circRNA array. An Agilent Scanner G2505C (Santa Clara, CA, USA) was used to scan the images, and the data were analyzed using Agilent Feature Extraction software (version 11.0.1.1) and R software, respectively. A heat map showing circRNA expression was generated based on the following criteria: fold change >2 and P<0.05.

Quantitative real-time polymerase chain reaction (qRT-PCR)

RNA was isolated with a TRIzol reagent (Sigma-Aldrich). RNA was reverse-transcribed to complementary DNA (cDNA) using a superscript reverse transcriptase (Invitrogen) and assessed using a ViiA7 quantitative PCR system (Applied Biosystems, Carlsbad, CA, USA). The change in RNA levels was calculated using the 2^−ΔΔCT method and normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH).

Ribonuclease (RNase) R digestion assay

RNase R digestion assays were used to evaluate the stability of circFTO in neonatal rat cardiomyocytes (NRCMs). Total RNAs (2 µg) were incubated briefly with 3 U/mg RNase R (Lucigen, Madison, WI, USA) at 37 ℃ for 30 min. The cells were harvested after 24 h of incubation. After treatment with RNase R, qRT-PCR assays were used to determine the expression of the circFTO and FTO mRNA.

Nuclear and cytoplasmic extraction

Nuclear and cytoplasmic fractions were isolated using the PARIS™ kit (Thermo Fisher, Waltham, MA, USA). Heart tissues or NRCMs were lysed for 10 min on ice with cell fractionation buffer. After centrifugation at 4 ℃ and 500 ×g for 3 min, the supernatants were collected as cytoplasmic fractions. Finally, the pellets were lysed with cell disruption buffer to obtain nuclear fractions. qRT-PCR assays were conducted to determine circFTO expression in the cytoplasmic and nuclear fractions.

Fluorescence in situ hybridization (FISH)

BioTech Co., Ltd. (Shanghai, China) designed and synthesized a probe for circFTO, covering the specific junction region of circFTO. To detect the subcellular location of circFTO, cardiomyocytes were fixed with 10% neutral buffered formalin on slides. The detection reagent kit by FISH (Genelily Biotech, Shanghai, China) was used to identify the signals of the circFTO probe. A Zeiss LSM800 confocal microscope (Zeiss, Heidelberg, Germany) was used to capture the images.

Isolation and culture of mouse neonatal ventricular myocytes

The standard enzymatic method was used to isolate the mouse neonatal left ventricular myocytes from the neonatal mouse hearts (45). The cells were then treated with 10⁻⁶ mol/L angiotensin-II (Ang-II) for 48 h before being collected for further analysis.

Construction and delivery of recombinant adenovirus-associated virus (AAV)

Recombinant adenovirus (Ad) and serum type 9 AAV expressing mouse FTO, circFTO, and IGF2BP2 short hairpin RNAs (shRNAs) were prepared using the pAd-Max and pAAV9 helper-free vector system (Genelily Biotech) (44). Primary cardiac myocytes were infected with Ad or AAV particles at a multiplicity of infection of 100. In the in vivo study, the surgical procedures and AAV delivery were performed as previously reported (46). Recombinant AAV9 vectors carrying shRNA against circFTO or control shRNA were injected via the tail vein at a titer of 1×10¹² vector genomes (vg) per mouse in a volume of 100 µL, 3 days post-TAC surgery.

Immunofluorescence analysis

Immunofluorescence assays were carried out (44). The cardiomyocyte cells fixed with paraformaldehyde were treated with 0.5% Triton X-100 in phosphate-buffered saline (PBS) for 15 min. The myocytes were then stained with anti-actinin (1:100) antibodies. After washing with PBS, fluorescein isothiocyanate-conjugated anti-rabbit immunoglobulin G (IgG; 1:2,000, Jackson, West Grove, PA, USA) was stained for 2 h. The stained cells were examined using a Zeiss LSM800 confocal microscope (Zeiss, Oberkochen, Germany).

Protein/DNA ratio detection

After washing with PBS, the cellular protein and DNA content of the cardiomyocytes were assessed. After being treated with 0.2 N perchloric acid (1 mL) and incubated with sodium hydroxide (250 mL, 0.3 N, 60 ℃, 20 min), the sample was centrifuged (10,000 g, 10 min). The protein content was analyzed using the Lowry method, with serum albumin as the standard. Salmon sperm DNA was used as the standard for analyzing DNA content using the Hoechst dye 33258 (MCE, Shanghai, China).

Western blot analysis

Radio immunoprecipitation assay (RIPA) cell lysis buffer (Beyotime Institute of Biotechnology, Nanjing, China) was used to quickly dissolve heart tissues or myocardial cells at 4 ℃. The protein samples were separated by 8–10% sodium dodecyl-sulfate polyacrylamide gel electrophoresis, transferred to nitrocellulose membranes (Millipore, Billerica, MA, USA), and blocked with 10% non-fat milk in tris-buffered saline and Tween. The membranes were incubated with primary antibody at 4 ℃ and then incubated with secondary antibodies for 1 h at room temperature. The signals were identified using an enhanced chemiluminescence kit (Thermo Scientific, Rockford, IL, USA) and analyzed by ImageQuant LAS4000 (GE, Chicago, IL, USA).

Mass spectrometry analysis

The sample proteins were examined by electrophoresis, and a specific band was removed. The digested proteins were subjected to the Orbitrap Velos Pro LC/MS system (Thermo Fisher Scientific, Waltham, MA, USA). The fragment spectra were used to check the National Center for Biotechnology Information (NCBI) database using the Mascot search engine.

Dual-luciferase reporter assays

The wild-type IRES was built into the luciferase reporter gene of the circFTO plasmid in the 293T cells. A dual-luciferase reporter test was conducted (Promega, Madison, WI, USA), and the test was performed in triplicate.

Polysome profiling analysis

The heart tissues or NRCMs were lysed in 500 µL of polysome lysis buffer {5 mM Tris-HCl (pH 7.5), 2.5 mM of MgCl₂, 1.5 mM of KCl, 1× protease inhibitor cocktail [ethylenediaminetetraacetic acid (EDTA)-free], 0.5% Triton X-100, 2 mM of dithiothreitol (DTT), 0.5% sodium deoxycholate, 100 units of RNase inhibitor, and 100 µg/mL of cycloheximide (CHX)} for 15 min at 4 ℃. Afterward, the samples were centrifuged at 4 ℃ for 7 min at 16,000 ×g to obtain the nuclei and mitochondria. The supernatant was loaded onto a 5–50% (w/v) sucrose density gradient, and ultracentrifuged at 20,000 ×g for 2 h at 4 ℃ using a Beckman SW41 rotor (Fullerton, CA, USA), then fractionated BioComp PGFip Piston Gradient Fractionator Model 152 (Biocomp, New Brunswick, Canada). The samples were measured using an absorbance detector connected to a fraction collector at a 254 nm wavelength. RNA was extracted from fractions using TRIzol LS solution (Invitrogen), and the circFTO levels were evaluated using the Northern blotting technique using Mupid^®-2 Mini-Gel System (Cosmo Bio Co., Seoul, Korea).

Total m⁶A methylation quantification

The m⁶A methylation quantification technique was used to calculate the total RNA from each sample. The test was performed using a total m⁶A methylation quantification kit (Colorimetric, Genelily Biotech).

Dot blot

Each sample of total RNA was heated to 75 ℃ for 5 min, left to cool for 1 min, and loaded onto Amersham Hybond-N+ membranes (0.45 µm, Solarbio, Beijing, China). The membranes were crosslinked, blocked, and incubated with an anti-m⁶A antibody (1:2,000, Abcam, Cambridge, MA, USA) at 4 ℃ overnight. The membranes were rinsed and probed with a horseradish peroxidase (HRP)-conjugated secondary anti-rabbit IgG at 25 ℃ for 1 h. The membranes were washed three times and treated with chemiluminescent HRP substrate (Millipore). After the treatment, the membranes were stained with methylene blue staining buffer using gentle shaking for 30 min, and then rinsed with ribonuclease-free water. Afterward, the input RNA was scanned to calculate its total content.

Co-immunoprecipitation (Co-IP) assays

Co-IP assays were performed using the Pierce Co-IP Kit (Thermo Fisher, Waltham, MA, USA). The cell lysates were incubated with AminoLink plus coupling resin immobilized primary antibody at 4 ℃ overnight. The samples were washed thrice with 200 µL of wash buffer and eluted with elution buffer for 5 min. Western blot assays were used to analyze the eluates.

Protein half-life analysis

The normal or FTO36aa-treated H9C2/NRCM cells were treated with CHX (20 µM, HY-12320, MedChemExpress, Shanghai, China), MG132 (10 µM, HY-13259, MedChemExpress), and chloroquine phosphate (10 µM, HY-17589A, MedChemExpress) at 0, 2, 4, and 8 h. The total proteins were prepared and calculated by Western blot assays.

Methylated RNA immunoprecipitation (Me-RIP)-qPCR

Me-RIP was performed using a RIP kit (Genelily Biotech). The cells were lysed with 1 mL of RIP lysis buffer for 10 min, and 100 µL of RIP lysis buffer was stored at −80 ℃. Anti-m⁶A antibodies (1:50) or normal rabbit IgG (1:50) were mixed with protein A/G beads at 4 ℃ for 2 h and incubated with 450 µL of the lysate at 4 ℃ for 2 h. After rinsing the beads with buffer, qRT-PCR was extracted and conducted.

RIP assay

The EZ-Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Millipore) was used for the RIP assays. A total of 100 µL of lysates of heart tissues (1 µg/mL) were incubated with 50 µL of magnetic beads combined with 5 µg of anti-IGF2BP1/2/3 at 4 ℃ overnight. Mouse IgG antibody (Cat# sc-2025, Santa Cruz, CA, USA) was used as the negative control. The immunoprecipitated RNA was isolated using proteinase K and analyzed by qRT-PCR assays.

Statistical analysis

The data are presented as the mean ± standard error of the mean. Multiple comparisons or repeated measures using analysis of variance or repeated analysis of variance were used to evaluate statistical significance. The Student t-test was used to compare the significant mean differences. Overall, significance was set at P<0.05.

Results

CircRNA circFTO was significantly elevated in hypertrophic cardiomyocytes

This study employed circRNA microarray screening to identify dysregulated circRNAs in the TAC mice and to examine their potential regulatory role in myocardial fibrosis. The heatmap analysis (Figure 1A) revealed significantly upregulated circRNAs (fold change >20) in the TAC mice compared to the sham mice. Most of the dysregulated circRNAs (which included 235 upregulated circRNAs and 84 downregulated circRNAs; fold change >2) were obtained from exonic regions (Figure 1B). Among the upregulated circRNAs, mmu_circ_0001717 (referred to as circFTO), derived from the critical m⁶A demethylase FTO protein, was generated from the back-splicing of exon 5 and exon 7 of the FTO gene (Figure 1C).

Figure 1.

CircRNA circFTO was significantly elevated in hypertrophic cardiomyocytes. (A) Heatmap showing the significantly upregulated circRNAs (with a fold change >20) in the TAC mice compared to the sham mice. (B) Pie chart showing the types of dysregulated circRNAs. (C) The selected circRNA (mmu_circ_0001717; termed circFTO) is derived from the critical m⁶A demethylase FTO protein, generated from the back-splicing of exon 5 and exon 7 from the FTO gene. (D) The upregulated circRNAs were verified by RT-PCR in the TAC mice and PE-treated hypertrophic NRCMs. (E) Divergent primers detected circFTO in cDNA but not in gDNA. (F) RT-PCR analysis of circFTO and linear FTO mRNA after treatment with RNase R in cardiomyocytes showed the RNase R digestion resistant ability of circFTO. (G) Subcellular distribution of circFTO was most abundant in the cytoplasm, as revealed by nuclear mass separation assays and (H) FISH. N=6. Compared to indicated group, *, P<0.05; ***, P<0.001. cDNA, complementary DNA; circFTO, circRNA derived from the FTO gene; circRNA, circular RNA; DAPI, 4',6-diamidino-2-phenylindole; FISH, fluorescence in situ hybridization; FTO, fat mass and obesity-associated; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; gDNA, genomic DNA; m⁶A, N6-methyladenosine; mRNA, messenger RNA; NRCM, neonatal rat cardiomyocyte; PE, phenylephrine; RNase, ribonuclease; RT-PCR, real-time polymerase chain reaction; TAC, transverse aortic constriction.

The RT-PCR analysis showed that circFTO upregulation was most notable in the phenylephrine (PE)-treated hypertrophic NRCMs (Figure 1D). Two sets of primers were designed to characterize the circular feature of circFTO, with divergent primers used to amplify circular transcripts, and convergent primers used to identify linear transcripts. Agarose gel electrophoresis showed that product bands were amplified from both cDNA and genomic DNA (gDNA) using convergent primers, but no product band was amplified from cDNA using divergent primers (Figure 1E). The RNase R degradation assays confirmed the stability of circFTO, which was resistant to RNase R digestion (Figure 1F). Cytoplasmic/nuclear fraction qRT-PCR (Figure 1G) and FISH analysis (Figure 1H) indicated that circFTO was mainly localized to the cytoplasm. In conclusion, the results suggest that circFTO is significantly increased in hypertrophic cardiomyocytes and may play a critical role in the regulation of myocardial fibrosis.

Silencing circFTO attenuates PE-induced neonatal cardiomyocytes hypertrophy in vitro

In this study, a cellular PE-induced hypertrophy model was established to explore the role of upregulated circFTO in cardiac hypertrophy. To suppress circFTO expression, recombinant Ad expressing shRNA targeting the junction sequence of circFTO was constructed (Figure 2A). RT-PCR was used to verify the suppression of circFTO expression. Both normal and subdue circFTO cardiomyocytes were treated with PE (50 µmol/L) for 48 h. The upregulated circFTO in the PE-treated NRCMs was significantly suppressed by (Ad-mediated expression of FTO shRNA) Ad-shcircFTO (Figure 2B).

Figure 2.

Silencing of circFTO attenuates PE-induced neonatal cardiomyocyte hypertrophy in vitro. (A) The specific siRNA sequence targeting the junction sequence. (B) The silencing efficiency of circFTO was verified by RT-PCR. (C) The effect of circFTO silencing on the hypertrophic phenotype of cardiomyocytes; the cell cross size, (D) protein/DNA ratio, (E) mRNA, and (F) protein expression levels of hypertrophic genes were detected. N=6. Compared to indicated group, *, P<0.05; **, P<0.01; ***, P<0.001. CircFTO, circRNA derived from the FTO gene; circRNA, circular RNA; DAPI, 4',6-diamidino-2-phenylindole; FTO, fat mass and obesity-associated; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; mRNA, messenger RNA; NC, negative control; PE, phenylephrine; RT-PCR, real-time polymerase chain reaction; sh, short hairpin; siRNA, small interfering RNA.

In terms of hypertrophic phenotype, both the cell cross size (Figure 2C) and protein/DNA ratio (Figure 2D) were significantly increased in the PE-treated NRCMs, indicating hypertrophic NRCMs. The knockdown of circFTO in the normal NRCMs did not affect the cell cross size and protein/DNA ratio. However, the knockdown of circFTO significantly decreased these indices in the PE-treated NRCMs, suggesting an attenuated hypertrophic phenotype. Moreover, the mRNA (Figure 2E) and protein (Figure 2F) expression levels of the hypertrophic genes, including ANP, BNP, and β-MHC, were suppressed by the knockdown of circFTO, indicating the hypertrophy-promotive role of circFTO.

Together, these results suggest that upregulated circFTO plays a critical role in promoting hypertrophic phenotype in cardiomyocytes, and the knockdown of circFTO may be a potential therapeutic strategy for inhibiting pathological myocardial hypertrophy.

Silencing circFTO attenuates pressure overload-induced cardiac hypertrophy in vivo

To validate the pro-hypertrophy effect of circFTO in vivo, a TAC-induced hypertrophy model was established. Four weeks after AAV expressing circFTO shRNA and TAC surgery, the level of circFTO in the left ventricular tissues was quantified by RT-PCR. The results showed that upregulated circFTO in the TAC mice was significantly repressed by AAV-shcircFTO (Figure 3A). Additionally, the knockdown of circFTO reduced the heart weight/body weight (HW/BW) of the TAC mice (Figure 3B).

Figure 3.

Silencing circFTO attenuates pressure overload-induced cardiac hypertrophy in vivo. A TAC-induced hypertrophy model was established. Four weeks after the delivery of shcircFTO expressing AAV to silence circFTO in the left ventricular tissues, (A) the efficiency of circFTO silence was verified by RT-PCR. (B) Silencing circFTO decreased TAC-increased HW/BW. (C) Histology analysis, including H&E staining (scale bar: 50 μm), WGA staining (scale bar: 20 μm), and Masson’s trichrome staining (scale bar: 50 μm), showed that circFTO silencing attenuated TAC-induced cardiomyocyte enlargement and collagen fibrosis. (D) CircFTO silencing decreased TAC impaired cardiac functions, including increased LVEF, LVFS, and decreased LVEDV, LVESV, LVEDD, LVESD, LVPWT, and IVS. (E,F) CircFTO silencing decreased the TAC mRNA and (G,H) protein levels of cardiac hypertrophy markers such as Myh7, ANP, BNP, and β-MHC, as well as myocardial fibrosis markers such as Postn, Spp1, fibronectin, collagen I, α-SMA, TGFB3, and tensin. N=6. Compared to indicated group, *, P<0.05; **, P<0.01; ***, P<0.001. AAV, adenovirus-associated virus; circFTO, circRNA derived from the FTO gene; circRNA, circular RNA; LVEF, left ventricular ejection fraction; LVFS, left ventricular fractional shortening; FTO, fat mass and obesity-associated; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; H&E, hematoxylin and eosin; HW/BW, heart weight/body weight; IVS, interventricular septum; LVEDV, left ventricular end-diastolic volume; LVEDD, left ventricular end-diastolic dimension; LVEF, left ventricular ejection fraction; LVFS, left ventricular fractional shortening; LVESD, left ventricular end-systolic dimension; LVESV, left ventricular end-systolic volume; LVPWT, left ventricular posterior wall thickness; mRNA, messenger RNA; NC, negative control; sh, short hairpin; TAC, transverse aortic constriction; WGA, wheat germ agglutinin.

Consistent with these findings, the circFTO-knockdown mice had fewer enlarged cardiomyocytes and less fibrosis as detected by WGA and Masson staining (Figure 3C). Moreover, cardiac echocardiography revealed that the knockdown of circFTO protected impaired heart function, with higher LVEDV, left ventricular end-systolic volume (LVESV), left ventricular end-diastolic dimension (LVEDD), LVESD, left ventricular posterior wall thickness (LVPWT), and interventricular septum (IVS), reducing LVEF and LVFS (Figure 3D).

As predicted, the knockdown of circFTO reduced the expression of cardiac hypertrophy markers such as Myh7, ANP, BNP, and β-MHC (Figure 3E,3F), and myocardial fibrosis markers such as Postn, Spp1, fibronectin, collagen I, α-SMA, TGFB3, and tensin in the TAC hearts (Figure 3G,3H). In summary, the results suggest that the knockdown of circFTO reduces pressure overload-induced cardiac hypertrophy in vivo. Therefore, circFTO may serve as a potential target for the prevention and treatment of cardiac hypertrophy and fibrosis.

FTO-36aa in cardiomyocytes

This study then aimed to examine the protein-coding ability of the cytoplasmic localized circFTO and explore its regulatory mechanism during myocardial hypertrophy. Using the NCBI ORF finder tool, circFTO was found to contain one ORF and likely encoded the 36 amino acid protein (Figure 4A). The liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis further confirmed the presence of a unique sequence of HSLGLVPLTVWQSAQ, consistent with the predicted sequence of FTO-36aa (Figure 4B).

Figure 4.

FTO-36aa in cardiomyocytes. (A) The NCBI ORF finder tool showed that circFTO contained one ORF and might encode a 36 amino acid protein. (B) The LC-MS/MS analysis showed the unique sequence of HSLGLVPLTVWQSAQ, which was consistent with the predicted sequence of FTO-36aa. (C) One vector control and three lentivirus-mediated overexpression vectors, including one full-length wild-type circFTO, one full-length circFTO with start codon mutant, and one full-length of circFTO-ORF, were constructed. After the infection of the NRCMs, the level of circFTO was significantly increased in the groups of lentiviruses containing full-length wild-type or mutant circFTO. (D) The flag-conjugated protein was detected in the groups of lentiviruses containing full-length wild-type circFTO or wild-type circFTO-ORF but not in the group of lentiviruses containing full-length mutant circFTO. (E,G) The RT-PCR and (F,H) Northern blotting of the sucrose gradient fractions showed an increased enrichment of circFTO in the polysome fraction of the TAC hearts and PE-treated NRCMs, which was inhibited by puromycin, a translation inhibitor. (I) Dual-luciferase analysis showed that Fluc/Rluc activity was significantly decreased when the wild-type IRES was deleted. (J) IF results detected a positive staining signal of FTO-36aa in NRCMs, further enhanced by PE treatment. N=3. Compared to indicated group, ***, P<0.001. CircFTO, circRNA derived from the FTO gene; circRNA, circular RNA; DAPI, 4',6-diamidino-2-phenylindole; FTO, fat mass and obesity-associated; FTO-36aa, circFTO-encoded a 36 amino acid protein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; IF, immunofluorescence; IRES, internal ribosome entry site; LC-MS/MS, liquid chromatography-tandem mass spectrometry; NCBI, National Center for Biotechnology Information; NRCM, neonatal rat cardiomyocyte; ORF, open reading frame; PE, phenylephrine; RT-PCR, real-time polymerase chain reaction; TAC, transverse aortic constriction.

To validate the coding ability of circFTO-ORF, three lentivirus-mediated overexpression vectors were constructed as displayed in the left panel of Figure 4C, including one full-length wild-type circFTO, one full-length circFTO with a start codon mutation, and one circFTO-ORF. All these vectors inserted three repeated flag sequences at the initial codon. After the infection of the NRCM cells, the circFTO level increased in the lentivirus groups containing full-length wild-type or mutant circFTO (Figure 4D). However, the flag-conjugated protein was only observed in the lentivirus groups containing full-length wild-type circFTO or wild-type circFTO-ORF, and was not observed in those containing full-length mutant circFTO (Figure 4D).

Sucrose gradients were then examined to separate the polysome-associated RNAs and analyze the translatable circFTO. Both the qPCR (Figure 4E) and Northern blotting (Figure 4F) tests showed an intense enrichment of circFTO in the polysome fraction of the TAC hearts and PE-treated NRCMs (Figure 4G). However, treatment with puromycin, a translation inhibitor, resulted in the transfer of circFTO to the light ribosome fraction (Figure 4G). The changes were further confirmed by the Northern blotting test (Figure 4H). Meanwhile, the circFTO-ORF was inserted into the dual-luciferase reporter to evaluate the IRES activity with or without the IRES element. The luciferase assay results showed that the Firefly luciferase/Renilla luciferase ratio (Fluc/Rluc) activity was significantly decreased when the wild-type IRES was removed (Figure 4I).

To further validate the presence of FTO-36aa, an antibody was created to target FTO-36aa (antigen sequence: AHSLGLVPLTVWQSAQQAPW). The immunofluorescence (IF) results detected a positive staining signal of FTO-36aa in the NRCMs, enhanced by PE treatment (Figure 4J). In summary, the results suggest that FTO-36aa in cardiomyocytes and its translation is regulated by an IRES-dependent mechanism.

FTO-36aa protein played a critical role in the pro-hypertrophy effect of circFTO

To investigate the role of circFTO-encoded FTO-36aa in the pro-hypertrophy effect, FTO-36aa was synthesized and injected via the tail vein of the TAC mice 3 days after surgery. Four weeks after TAC surgery, the level of circFTO in the left ventricular tissues was monitored by RT-PCR. The results showed that circFTO was still suppressed in the circFTO-silenced mice injected with FTO-36aa (Figure 5A). FTO-36aa did not affect the heart morphology parameters, including the HW/BW (Figure 5B), cardiomyocyte size, and collagen deposition (Figure 5C). However, the hypertrophic effects induced by circFTO suppression, including the increased HW/BW, cardiomyocyte enlargement, and fibrosis, were almost completely reversed by FTO-36aa.

Figure 5.

FTO-36aa protein played a critical role in the pro-hypertrophy effect of circFTO. Recombinant FTO-36aa protein was injected via the tail vein in the TAC mice 3 days after surgery. (A) Four weeks after TAC surgery, the suppressed level of circFTO in the left ventricular tissues was quantified by RT-PCR, which was unchanged by the FTO-36aa injection. (B) FTO-36aa increased circFTO silencing and decreased the HW/BW of the TAC mice. (C) The histology analysis, including H&E staining (scale bar: 500 µm and 50 μm), WGA staining (scale bar: 20 μm), and Masson’s trichrome staining (scale bar: 50 μm), showed that FTO-36aa increased circFTO silencing, and decreased cardiomyocyte enlargement and collagen fibrosis. (D) FTO-36aa impaired circFTO silencing and protected cardiac functions, including decreased LVEF and LVFS, and increased LVEDV, LVESV, LVEDD, LVESD, LVPWT, and IVS. (E,F) FTO-36aa increased circFTO silencing and suppressed the mRNA and (G,H) protein levels of cardiac hypertrophy markers such as Myh7, ANP, BNP, and β-MHC, as well as myocardial fibrosis markers such as Postn, Spp1, fibronectin, collagen I, α-SMA, TGFB3, and tensin. N=6. Compared to indicated group, *, P<0.05; **, P<0.01; ***, P<0.001. CircFTO, circRNA derived from the FTO gene; circRNA, circular RNA; LVEF, left ventricular ejection fraction; LVFS, left ventricular fractional shortening; FTO, fat mass and obesity-associated; FTO-36aa, circFTO-encoded a 36 amino acid protein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; H&E, hematoxylin and eosin; HW/BW, heart weight/body weight; IVS, interventricular septum; LVEDV, left ventricular end-diastolic volume; LVEDD, left ventricular end-diastolic dimension; LVEF, left ventricular ejection fraction; LVFS, left ventricular fractional shortening; LVESD, left ventricular end-systolic dimension; LVESV, left ventricular end-systolic volume; LVPWT, left ventricular posterior wall thickness; mRNA, messenger RNA; NC, negative control; sh, short hairpin; TAC, transverse aortic constriction; WGA, wheat germ agglutinin.

Cardiac echocardiography also indicated that circFTO silencing improved heart function, as evidenced by decreased LVEDV, LVESV, LVEDD, LVESD, LVPWT, and IVS, and increased LVEF and LVFS, which were comprehensively reversed by the FTO-36aa injection (Figure 5D).

The RT-PCR and Western blotting results further revealed that circFTO suppression reduced the expression of cardiac hypertrophy markers (Myh7, ANP, BNP, and β-MHC) (Figure 5E,5F) and myocardial fibrosis indicators (Postn, Spp1, fibronectin, collagen I, α-SMA, TGFB3, and tensin) (Figure 5G,5H) in the TAC hearts, which were partially restored by the FTO-36aa injection (Figure 5D). These findings indicate that circFTO plays a role in promoting hypertrophy by encoding FTO-36aa, which may represent a potential therapeutic target for cardiac hypertrophy and fibrosis.

FTO-36 upregulated global m6A methylation by the ubiquitination-mediated protein degradation of FTO

The presented study investigated the downstream mechanism of FTO-36aa in the progression of myocardial hypertrophy. Report has shown that FTO-dependent m⁶A regulation contributes to remodeling and repair processes (43). Therefore, we examined global m⁶A methylation in heart tissues with or without circFTO knockdown or FTO-36aa injection. The results showed that global m⁶A methylation was reduced in the TAC mice via circFTO knockdown (Figure 6A). Interestingly, the FTO-36aa injection increased global m⁶A methylation in both the sham and TAC mice (Figure 6A), which was further validated by the dot blot analysis (Figure 6B).

Figure 6.

FTO-36aa upregulated the global m⁶A methylation by the ubiquitination-mediated protein degradation of FTO. (A) The global m⁶A methylation was significantly increased in the TAC mice and decreased by circFTO silencing. The FTO-36aa injection significantly increased the global m⁶A methylation in both the sham and TAC mice. (B) These changes were verified by dot blot. (C) The mRNA and (D) protein levels of FTO in the TAC mice with circFTO silencing or FTO-36aa injection were detected by RT-PCR and Western blotting. (E) The promoter activity assay showed that FTO-36aa did not affect the promoter activity of FTO. (F) The total ubiquitin analysis showed that FTO-36aa treatment promoted the ubiquitin levels of FTO. (G) Co-IP with specifical antibodies that recognize ubiquitin or (H) ubiquitin ligase E showed an increased FTO protein level after FTO-36aa treatment. (I) CHX (50 mg/mL) was used to inhibit new protein synthesis, and FTO-36aa promoted FTO protein degradation and shortened its half-life in the H9C2 and (J) NRCM cells. (K) Proteasome inhibitors MG132 (10 mM), bortezomib (50 mM), and chloroquine (10 mM) significantly inhibited FTO-36aa and promoted FTO protein degradation in the H9C2 and (L) NRCM cells. N=3. Compared to indicated group, *, P<0.05; **, P<0.01; ***, P<0.001. CHX, cycloheximide; circFTO, circRNA derived from the FTO gene; circRNA, circular RNA; co-IP, co-co-immunoprecipitation; FTO, fat mass and obesity-associated; FTO-36aa, circFTO-encoded a 36 amino acid protein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; IB, immunoblotting; IP, immunoprecipitation; m⁶A, N6-methyladenosine; mRNA, messenger RNA; NRCM, neonatal rat cardiomyocyte; RT-PCR, real-time polymerase chain reaction; sh, short hairpin; TAC, transverse aortic constriction.

Further, we found that FTO mRNA expression was significantly decreased in the TAC mice, and this effect was partially reversed by circFTO knockdown or FTO-36aa injection (Figure 6C). Surprisingly, the FTO protein levels were increased by FTO-36aa in the sham mice but decreased in the TAC mice, and this effect was reversed by circFTO knockdown (Figure 6D). Additionally, FTO-36aa had no effect on FTO promoter activity (Figure 6E), indicating that FTO expression was regulated by post-translational mechanisms mediated by FTO-36aa.

We further identified a total ubiquitin level of FTO-36aa in whole NRCM cell lysate, and found that the FTO-36aa treatment promoted the ubiquitin levels of FTO (Figure 6F). To validate the increased ubiquitin level of FTO, H9C2, and NRCM cells were treated with higher dosages of FTO-36aa, and immunoprecipitation with antibodies specific to recognizing ubiquitin (Figure 6G) or ubiquitin ligase E3 SKP2 (Figure 6H) was performed. The results showed that the FTO protein level increased, indicating a dose-dependent increase in ubiquitin levels of FTO mediated by FTO-36aa. Additionally, using CHX (50 mg/mL) to inhibit new protein synthesis, we found that FTO-36aa induced FTO protein degradation and shortened its half-life in the H9C2 cells (Figure 6I) and NRCMs (Figure 6J).

Moreover, the proteasome inhibitor MG132 (10 mM), bortezomib (50 mM), and chloroquine (10 mM) significantly hampered the FTO-36aa-promoted FTO protein degradation in the H9C2 (Figure 6K) and NRCM (Figure 6L) cells. These findings suggest that FTO-36aa promotes FTO protein degradation through a ubiquitin-proteasome system-dependent pathway, increasing global m⁶A methylation in cardiomyocytes.

m⁶A reader IGF2BP2 mediated circFTO/FTO-36aa regulated mRNA levels of hypertrophic genes

Previous study has shown that FTO demethylation regulates several contractile transcripts in cardiomyocytes, including Nppa, Myh7, Myh9, Serca2, Ryr2, and Ttn, as well as the myocardium-specific long non-coding RNAs (lncRNAs) Mhrt and Chast (43). In our study, Me-RIP-qPCR was performed to detect the regulation of methylated levels by circFTO or FTO-36aa in the TAC mice. The results showed that the methylated levels were suppressed in the TAC mice with circFTO knockdown, whereas the FTO-36aa injection dramatically increased the methylated levels of these mRNAs in the context of circFTO knockdown (Figure 7A).

Figure 7.

The m⁶A reader IGF2BP2 mediated the circFTO/FTO-36aa regulated mRNA levels of the hypertrophic genes. (A) The Me-RIP-qPCR results revealed that the methylated levels of hypertrophic genes, including Nppa, Myh7, Myh9, Serca2, Ryr2, and Ttn, as well as myocardium-specific lncRNAs Mhrt and Chast, were significantly increased in TAC group and were suppressed by circFTO silencing. The FTO-36aa injection dramatically reversed the suppressed methylated of these mRNAs by circFTO silencing. (B) RIP with specific antibodies of IGF2BP1/2/3 was performed, and the enrichment of the hypertrophic mRNAs was analyzed by RT-PCR. (C) IGF2BP2 knockdown reversed both the PE or FTO-36aa elevated expression of these hypertrophic genes in the H9C2 and (D) NRCMs cells. N=3. Compared to indicated group, *, P<0.05; **, P<0.01; ***, P<0.001. CircFTO, circRNA derived from the FTO gene; circRNA, circular RNA; FTO, fat mass and obesity-associated; FTO-36aa, circFTO-encoded a 36 amino acid protein; lncRNA, long non-coding RNA; m⁶A, N6-methyladenosine; mRNA, messenger RNA; NC, negative control; NRCM, neonatal rat cardiomyocyte; PE, phenylephrine; qPCR, quantitative polymerase chain reaction; RIP, RNA immunoprecipitation; RT-PCR, real-time polymerase chain reaction; sh, short hairpin; TAC, transverse aortic constriction.

Considering the m⁶A methylation of these mRNAs and their relationship with increased mRNA levels, we hypothesized that the upregulation of these mRNAs could result from improved stability mediated by the IGF2BP m⁶A reader family (IGF2BP1, IGF2BP2, and IGF2BP3). Therefore, RIP assays were performed with specific antibodies against IGF2BP1/2/3. The results showed that the enrichment of the hypertrophic mRNAs was similar in the anti-IGF2PB1 antibody RIP complex between the circFTO-knockdown and FTO-36aa-treated TAC mice (Figure 7B). However, their enrichment in the anti-IGF2PB2 antibody RIP complex was significantly increased in the normal or FTO-36aa-treated TAC mice, which was partially impaired by circFTO knockdown (Figure 7B). Similar results were observed in the anti-IGF2PB3 antibody RIP complex, although to a lesser extent (Figure 7B), indicating that IGF2BP2 was the major m⁶A reader regulating the mRNA stability of hypertrophic genes.

To validate this hypothesis, we subdued IGF2PB2 in the H9C2 and NRCMs cells. The expression of the hypertrophic genes was significantly increased by the PE or FTO-36aa treatment (Figure 7C). The knockdown of circFTO slightly reduced the increment levels of these hypertrophic genes induced by PE but did not affect those induced by FTO-36aa (Figure 7C). Notably, the knockdown of IGF2BP2 largely reversed the elevated expression of these hypertrophic genes induced by PE or FTO-36aa in the H9C2 (Figure 7C) and NRCM (Figure 7D) cells. These findings suggest that the m⁶A reader IGF2BP2 mediates the circFTO/FTO-36aa-regulated mRNA levels of hypertrophic genes.

Role of IGF2BP2 in FTO-36aa/FTO in regulating m6A signaling in myocardial hypertrophy

To further confirm the role of IGF2BP2 in promoting the FTO-36aa/FTO-mediated regulation of m⁶A methylation and myocardial hypertrophy, we overexpressed FTO and subdued IGF2BP2 in the TAC mice with or without FTO-36aa treatment using AAV. The efficiency of FTO overexpression and IGF2BP2 inhibition was validated by RT-PCR (Figure 8A) and Western blotting (Figure 8B). FTO overexpression weakened the TAC-induced hypertrophic phenotype, as evidenced by the decreased HW/BW (Figure 8C), HW/tibia length (TL) (Figure 8D), cardiomyocyte enlargement, and collagen deposition (Figure 8E), and also reduced the mRNA (Figure 8F,8G) and protein (Figure 8H,8I) levels of the cardiac hypertrophy markers Myh7, ANP, BNP, and β-MHC, and the myocardial fibrosis markers Postn, Spp1, fibronectin, collagen I, α-SMA, TGFB3, and tensin. These beneficial changes were reversed by the FTO-36aa treatment. Moreover, subduing IGF2BP2 prevented the pro-hypertrophy effect of FTO-36aa.

Figure 8.

IGF2BP2 was the major m⁶A reader of FTO-36aa/FTO-regulated m⁶A signaling in myocardial hypertrophy progression. We overexpressed FTO and knocked down IGF2BP2 in the TAC mice with or without FTO-36aa treatment by AAV. (A) The efficiency of FTO overexpression and IGF2BP2 knockdown was confirmed by RT-PCR and (B) Western blotting. (C-I) Histology analysis, including H&E staining (scale bar: 50 μm), WGA staining (scale bar: 10 μm), and Masson’s trichrome staining (scale bar: 25 μm) showed the effect of FTO overexpression, FTO-36aa treatment, and IGF2BP2 knockdown on the TAC-induced hypertrophic phenotype, included decreased (C) HW/BW, (D) HW/TL, (E) cardiomyocyte enlargement and collagen deposition, (F,G) mRNA and (H,I) protein levels of cardiac hypertrophy markers (Myh7, ANP, BNP, and β-MHC) and myocardial fibrosis markers (Postn, Spp1, fibronectin, collagen I, α-SMA, TGFB3, and tensin) were analyzed. (J) The effects of FTO overexpression, FTO-36aa treatment, and IGF2BP2 knockdown on cardiac echocardiography were evaluated, including LVEDV, LVESV, LVEDD, LVESD, LVPWT, and IVS, as well as LVEF and LVFS. N=6. Compared to indicated group, *, P<0.05; **, P<0.01; ***, P<0.001. AAV, adenovirus-associated virus; circFTO, circRNA derived from the FTO gene; circRNA, circular RNA; LVEF, left ventricular ejection fraction; LVFS, left ventricular fractional shortening; FTO, fat mass and obesity-associated; FTO-36aa, circFTO-encoded a 36 amino acid protein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; H&E, hematoxylin and eosin; HW/BW, heart weight/body weight; IVS, interventricular septum; LVEDV, left ventricular end-diastolic volume; LVEDD, left ventricular end-diastolic dimension; LVEF, left ventricular ejection fraction; LVFS, left ventricular fractional shortening; LVESD, left ventricular end-systolic dimension; LVESV, left ventricular end-systolic volume; LVPWT, left ventricular posterior wall thickness; mRNA, messenger RNA; NC, negative control; sh, short hairpin; TAC, transverse aortic constriction; WGA, wheat germ agglutinin.

Cardiac echocardiography showed that FTO overexpression improved heart function by reducing LVEDV, LVESV, LVEDD, LVESD, LVPWT, and IVS, followed by increased LVEF and LVFS (Figure 8J). However, these beneficial cardiac contractile functions were suppressed by the FTO-36aa treatment (Figure 8J). With the knockdown of IGF2BP2, FTO-36aa did not impair these FTO-protected cardiac contractile functions. Overall, our results demonstrate that IGF2BP2 is the major m⁶A reader that mediates FTO-36aa/FTO-regulated m⁶A signaling in myocardial hypertrophy.

Discussion

Recent studies have shown that non-coding RNA aids in the occurrence and development of myocardial hypertrophy and heart failure (23,47,48). CircRNA is a large class of covalently closed circular non-coding RNAs (25). Unlike traditional linear RNA molecules, circRNAs have no 5' terminal cap or 3' poly(A) tail. CircRNAs are highly stable, unaffected by exonuclide, and widely abundant in eukaryotic cells (9). A study has shown that circRNAs play a crucial role in myocardial hypertrophy (48). Werfel et al. reported an increased cardiac expression of circRNAs in a TAC-induced mouse model of cardiac hypertrophy, which indicates that circRNAs may be associated with cardiac hypertrophy (49). Using both TAC- and Ang-II-induced hypertrophy mouse models, Xu et al. showed that circHIPK3 regulates cardiac hypertrophy (16). Lim et al. reported that circSlc8a1 acts as a sponge for miR-133a, inhibiting cardiac hypertrophy (18). Wang et al. found that heart-related circRNA, which serves as an endogenous miR-223 sponge, hampered hypertrophy induced by isoproterenol and TAC (17). Other circRNAs, including circACR, circCDYL, circNFIB, circYAP, and circNfix, have been linked to myocardial hypertrophy and ventricular remodeling (23,48,50). The present study found a significantly upregulated circRNA circFTO generated from the alternative back-splicing FTO pre-mRNA.

MiRNAs play essential roles in cellular homeostasis; thus, it is likely that some circRNAs regulate stress response pathways by inhibiting miRNA activity (51). CircRNAs can produce biological effects by interacting with specific proteins (52). Although most circRNAs cannot encode proteins, those that contain an IRES or m⁶A modification can be translated into small peptides or proteins (53). Legnini et al. recently proposed that human circ-ZNF609 is crucial in modulating myoblast proliferation and can be translated into peptides in a splicing-dependent and cap-independent manner (54). Other circRNAs, such as circMAP3K4 (55), circMAPK1 (29), circFNDC3B (30), circHEART5B (32), circHNRNPU (33), circSMO (34), and circAXIN1 (31), possess coding potential and exert biological functions in the development of various diseases, especially in tumors. However, no novel protein encoded by circular in cardiomyocytes was reported. Given the cytoplasmic localization and the translation regulatory elements of circFTO, it is likely that it has the potential to encode a novel protein. This study initially noted that circFTO was translatable and encoded a novel protein FTO-36aa from the junction ORF.

As a m⁶A demethylase, FTO aids in cardiac homeostasis, remodeling, and regeneration under stress states such as ischemia and hypoxia (56). The reduced expression of FTO in mammals with hypoxic cardiomyocytes increases the degree of m⁶A methylation in RNA, reducing the contractile function of cardiomyocytes (43). Improving FTO expression can reduce the m⁶A level and systolic myocardial function induced by ischemia (43). This can be achieved through the demethylation activity of FTO, which demethylates contractile-related transcripts, preventing its degradation and improving myocardial contractile protein expression during ischemia and myocardial contractility (43). FTO-dependent m⁶A methylation plays a major role in cardiac contraction in heart failure; thus, a therapy could potentially be developed using FTO or FTO analogs. Further, FTO-36aa enhances the ubiquitination-mediated protein degradation of FTO, weakening the demethylation of RNA and elevating the global m⁶A methylation of hypertrophic cardiomyocytes.

The m⁶A reader proteins can identify the m⁶A-modified RNAs divided into different protein families (57). One class of direct m⁶A reader proteins contains YTH domain 4, and several heterogeneous nuclear ribonucleoproteins regulate the alternative splicing or processing of target transcripts (58). IGF2BP1/2/3 and eukaryotic initiation factor 3 are from other subfamily members. Different species have individual m⁶A reader proteins for conducting specific biological functions (57). The m⁶A reader YTHDF2 suppresses cardiac hypertrophy via Myh7 mRNA decoy in a m⁶A-dependent manner (59). IGF2BP2 was shown to strengthen miR-133a binding and repress miR-133a’s targets during heart development and hypertrophy (60). In cardiac-specific FTO-subdued mice, a drastic reduction in EF and an increase in dilatation on TAC surgery were observed, increasing the methylation and expression of hypertrophic genes. The IGF2BP family likely promotes mRNA stability (43). This study showed that the m⁶A reader IGF2BP2 recognized the circFTO/FTO-36aa, increasing the m⁶A methylation and mRNA stabilities of m⁶A-methylated hypertrophic genes. Our data demonstrate that FTO-36aa-mediated degradation of FTO elevates global m6A levels, enabling IGF2BP2 to stabilize hypertrophy-related mRNAs (e.g., ANP, BNP, β-MHC). This axis directly links m6A-driven mRNA stability to ventricular remodeling.

This study had some limitations. It relied exclusively on a murine TAC model, which mimics pressure overload-induced hypertrophy but does not fully recapitulate the multifactorial etiology of human cardiac hypertrophy. The upstream factors regulating circFTO biogenesis in response to TAC were not investigated, limiting understanding of its dysregulation in disease. The study focused narrowly on m⁶A methylation, overlooking potential cross-talk with other RNA modifications or epigenetic mechanisms in hypertrophy.

Overall, protein-coding circRNA plays a role in regulating FTO-mediated m⁶A demethylation and contributes to pathological cardiac hypertrophy. As Figure 9 shows, the upregulated circRNA circFTO was generated from the back-splicing of FTO pre-mRNA between exon 5 and exon 7. A novel protein was identified and coded by circFTO, promoting the ubiquitination-mediated protein degradation of FTO, diminishing the demethylation of RNA, and increasing the global m⁶A methylation. The m⁶A reader IGF2BP2 stabilized the mRNA of m⁶A methylated hypertrophic genes. This study highlights the functional significance of the alternative splicing-generated circFTO and its coded FTO-36aa during cardiac hypertrophy. The outcome of this study provides a novel mechanistic insight into the m⁶A methylation regulation mechanisms in hypertrophic cardiomyocytes.

Figure 9.

Schematic diagram showing that an upregulated circRNA circFTO was generated from the alternative back-splicing of FTO pre-mRNA between exon 5 and exon 7. The silencing of circFTO by AAV significantly weakened the TAC-induced hypertrophy phenotype. A novel protein FTO-36aa was identified that was coded by circFTO, causing the pro-hypertrophy effect of the circuit. FTO-36aa promoted the ubiquitination-mediated protein degradation of FTO, suppressing the demethylation of RNA and elevating the global m⁶A methylation, which contributes to the upregulation of hypertrophic genes and the progression of myocardial hypertrophy. AAV, adenovirus-associated virus; circFTO, circRNA derived from the FTO gene; circRNA, circular RNA; FTO, fat mass and obesity-associated; FTO-36aa, circFTO-encoded a 36 amino acid protein; m⁶A, N6-methyladenosine; mRNA, messenger RNA; TAC, transverse aortic constriction.

Clinical perspectives

This study showed that pathological cardiac hypertrophy poses a significant threat to human health. CircRNAs have been shown to participate in myocardial hypertrophy and ventricular remodeling. Further, this study showed that alternative splicing-generated circFTO and its coded FTO-36aa were involved in myocardial hypertrophy. FTO-36aa promoted the ubiquitination-mediated protein degradation of FTO, which suppressed the demethylation of RNA, elevating global m⁶A methylation. These findings provide valuable insights into the molecular mechanisms underlying the m⁶A methylation regulation of hypertrophic cardiomyocytes. From a clinical perspective, targeting the circFTO/FTO-36aa-m⁶A-mRNA signaling pathway may hold promise in the development of therapeutic strategies to prevent or treat myocardial hypertrophy. Silencing circFTO or blocking FTO-36aa attenuated hypertrophy in mice, suggesting that peptide inhibitors of FTO-36aa or circFTO-targeted oligonucleotides could be developed for clinical use. However, human validation is essential; while we identified circFTO upregulation in murine models, future studies must assess its expression in human hypertrophic cardiomyopathy or heart failure tissues. Additionally, this study highlights the emerging significance of circRNAs in understanding the pathogenesis of various diseases, opening new avenues for further research in this field.

Conclusions

This study elucidated a novel mechanism by which the circRNA circFTO drives pathological cardiac hypertrophy through its encoded protein FTO-36aa, revealing a critical link between circRNA translation, m⁶A epigenetics, and myocardial remodeling. CircFTO, derived from the back-splicing of the FTO gene, is significantly upregulated in pressure-overloaded hearts and encodes a 36-amino acid protein that promotes the ubiquitination-mediated degradation of FTO. This degradation suppresses the demethylase activity of FTO, leading to elevated global m⁶A RNA methylation. The m⁶A reader IGF2BP2 recognizes these methylated transcripts, stabilizing the mRNAs of hypertrophic genes (e.g., ANP, BNP, and β-MHC), and exacerbating cardiomyocyte enlargement and fibrosis. Silencing circFTO or blocking FTO-36aa attenuates hypertrophy in vivo, restoring cardiac function and reducing pathological remodeling. These findings advance understanding of circRNA functionality beyond canonical roles as miRNA sponges, demonstrating their capacity to encode pathogenic proteins that modulate epigenetic landscapes. Clinically, the study underscores the therapeutic potential of targeting the circFTO/FTO-36aa/IGF2BP2 axis to mitigate m⁶A-driven hypertrophy, offering a paradigm shift in treating ventricular remodeling and heart failure. Future studies should be conducted to validate this pathway in human disease and explore the upstream regulators of circFTO biogenesis.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All the animal procedures followed National Institutes of Health (NIH) guidelines for the care and use of animals and were approved by the Animal Care and Use Committee of The First Affiliated Hospital, Nanchang University (No. 2020-62).

Data Sharing Statement

Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1462/dss

jtd-17-08-6127-dss.pdf ^{(159.4KB, pdf)}

DOI: 10.21037/jtd-2025-1462