1. Introduction
Heart failure (HF), a life-threatening syndrome arising from diverse pathologies, represents a major clinical challenge due to its elevated rates of morbidity and mortality. Pathological cardiac hypertrophy can cause an irreversible transition to HF or sudden death. In response to elevated blood pressure or the activation of neurohormones, cardiac hypertrophy is characterized by inflammation, collagen deposition, and changes in cardiomyocyte (CM) morphology and function. Many studies have revealed the molecular mechanisms involved in the various pathological processes of cardiac hypertrophy [1,
2]. However, effective therapeutic targets in the development of myocardial hypertrophy and HF are lacking.
Epigenetic modifications dynamically regulate multiple genomic processes, including gene expression, stability, localization, and function, and are implicated in HF and cardiac hypertrophy [3,
4]. Among them, N6-methyladenosine (m6A) RNA modification was first thought to be involved in cell proliferation, metabolism, apoptosis, and cell differentiation in various medical conditions, including cancer and heart [5,
6]. RNA 5-methylcytosine (m5C) is a prevalent RNA modification that can be found in transfer RNAs (tRNAs), ribosomal RNAs (rRNAs), messenger RNAs (mRNAs), and noncoding RNAs (ncRNAs) [7]. m5C modification involves the transfer of a methyl group to the C-5 position of the RNA cytosine base and is controlled by methyltransferases and demethylases [8]. The m5C modification regulates RNA metabolism, including mRNA stability, translation, and nuclear export [9], which are attributed to the recognition by the m5C reader proteins Aly/REF export factor (ALYREF) and Y-box binding protein 1 (YBX1), respectively [
10]. m5C participates in a range of pathophysiological processes, including tumor formation [
11].
The NOL1/NOP2/SUN domain (NSUN) family contains seven members: NSUN1–7, of which NSUN1, 2, and 5 isoforms are highly conserved [7]. Each family member catalyzes m5C modification in different RNA species [
12]. NOL1/NOP2/Sun domain family member 2 (NSUN2) is one of the predominant methyltransferase isoforms and causes m5C modifications in mRNAs, micro RNAs (miRNAs), long noncoding RNAs (lncRNAs) and tRNAs [13]. Numerous studies have indicated the importance of NSUN2-mediated m5C modifications in organ formation, senescence control, mitochondrial function maintenance, and cell growth [
14,
15]. NSUN2 plays a detrimental role in atherosclerosis, and its deficiency protects the endothelium from inflammation by modulating the translation efficiency of intercellular adhesion molecule-1 (ICAM-1) [16]. This highlights the possibility that NSUN2 and m5C modifications hold promise as novel biomarkers and therapeutic targets for cardiac remodeling and HF. Herein, we provide evidence that NSUN2 expression and RNA m5C levels are high in the myocardium of patients with HF and mouse hypertrophic hearts. Through gain- and loss-of-function experiments, we demonstrate that NSUN2 is a central regulator of pathological cardiac hypertrophy and HF. By recognizing the m5C-methylated modification sites of La-related protein 1 (LARP1) mediated by NSUN2, YBX1 enhances the stability of LARP1 and further binds to and prevents GATA-binding protein 4 (GATA4) mRNA degradation, thereby triggering a pro-hypertrophic phenotype. Collectively, we uncovered a previously unrecognized role and mechanism of NSUN2/YBX1/LARP1/GATA4 in cardiac hypertrophy. These data underscore NSUN2 as a promising therapeutic target for the development of preventative strategies against pathological cardiac hypertrophy and HF.
2. Material and methods
2.1. Ethics
Heart lysates from failing and non-failing hearts of patients were obtained from the tissue bank of the Heilongjiang Academy of Medical Sciences (China). The demographic characteristics of the human subjects are presented in Table S1 in Appendix A. The utilization of human cardiac tissues in this study was approved by the Ethics Committee of Harbin Medical University (No. HMUIRB20170034), and all experimental protocols strictly adhered to the ethical guidelines for human tissue research outlined in the Declaration of Helsinki.
Male C57BL/6 mice were obtained from Changsheng Biotechnology (China) and the animal center of the Second Affiliated Hospital of Harbin Medical University. The protocols for animal experiments were approved by the Ethic Committees of Harbin Medical University.
2.2. Mouse models
NSUN2 floxed mice (NSUN2flox/flox) were crossed with Myh6-Cre mice to generate NSUN2-cKO mice and littermate NSUN2flox/flox controls (C001041; Cyagen Biosciences, USA). For in vivo cardiac gene transfer, adeno-associated virus serotype 9 (AAV9) was employed to deliver constructs containing the cardiac troponin T (cTNT) promoter to ensure cardiac-specific expression. C57BL/6 mice were randomly assigned to receive either the adeno-associated virus serotype 9 (AAV9) virus that carries a specific plasmid for overexpressing NSUN2 (AAV9‐cTnT‐NSUN2-3 × flag, AAV9-NSUN2) or a negative control plasmid (AAV9‐cTnT‐Null, AAV9-Null) through tail vein injection. AAV9 carrying a short fragment for specific knockdown of LARP1 (AAV9‐cTnT‐shLARP1, shLARP1), a negative control scrambled RNA fragment (AAV9‐cTnT‐shNC, AAV9-shNC). The virus (1 × 1011 plaque-forming units per animal) was administered to the mice by injecting them into the tail vein.
Male mice were anaesthetized with avertin (Sigma-Aldrich, USA) and placed on a temperature-controlled surgical table for transverse aortic constriction (TAC) surgery. The transverse aorta was exposed by incising the sternum to the first intercostal, followed by ligation of the artery with a 5-0 wire between the first and second branches of the aortic arch.
Angiotensin II (Ang II, 2.5 mg∙kg–1∙d–1; Sigma-Aldrich) was delivered to mice by subcutaneously implanted Alzet osmotic mini-pumps (Model 2004) for four weeks. The control mice had a sterile incision made in their back to implant the mini-pump, and were provided with saline.
2.3. Tail-cuff blood pressure
Non-invasive blood pressure was measured in the tails of mice using Volume Pressure Recording (VPR) sensor technology (Kent Scientific, USA). Briefly, mice were secured in tubes and placed on a warming pad after equipment leak detection. The tails were exposed through the VPR sensor, and data were collected using the Kent software. The program consisted of a five-cycle check phase and ten sampling cycles, with a 10 s interval between each cycle.
2.4. Echocardiographic analyses
Cardiac functions were evaluated using a Vevo2100 echocardiography system (VisualSonics, Canada) equipped with a 10 MHz phase-array transducer. Mice were anaesthetized with avertin and positioned on a heating pad maintained at 37 °C.
2.5. Histological analyses
Mouse hearts were dissected out and fixed in paraformaldehyde (PFA; Biosharp, China), then sectioned into 4-μm slices. These sections were embedded in paraffin for subsequent analyses. Histological analyses employing Masson’s trichrome and wheat germ agglutinin (WGA; Invitrogen, USA) assessed myocardial architecture and collagen deposition in serial heart sections to evaluate structural remodeling. Images were analyzed by ImageJ or Image-Pro Plus.
2.6. Immunofluorescence
Specimens (cells or cardiac tissue) underwent PFA-mediated immobilization following a standardized 20 min fixation protocol to preserve ultrastructural integrity prior to downstream analyses. To permeabilize the cells, a solution comprising 10 mL phosphate-buffered saline (PBS), 10 μL Triton-100, and 0.01 g sodium citrate was applied for 1 h, followed by blocking with goat serum for blotting procedures. An anti-α-actinin antibody, recognizing a cytoskeletal protein, was utilized as a marker for CMs. Immunofluorescence was quantified at 40× magnification using FV300 microscope from Olympus (Japan), and cell surfaces were measured by ImageJ software.
2.7. Plasmid construction
The plasmids for wild-type LARP1 (LARP1-WT, NM_028451) and mutant LARP1 (LARP1-Mut, NM_028451) were constructed as follows. The complementary DNAs (cDNAs) for both LARP1-WT and LARP1-Mut were purified by GENECHEM (China). These cDNAs were then cloned into the GV712 vector (cytomegalovirus (CMV) enhancer-multiple cloning site (MCS)-simian vacuolating virus 40 (SV40)-puromycin) to generate the overexpression constructs. In the LARP1-Mut construct, three cytosine (C) residues located within the region spanning chr11:58055466 to 58055480 were substituted with thymidine (T) to abolish the methylation potential of the gene. All inserted sequences were confirmed by DNA sequencing prior to further use.
2.8. Cell culture
Primary neonatal mouse CMs were isolated from 1–3-day-old C57BL/6 mice following established protocols. The harvested cells were cultured in basic 1× Dulbecco’s Modified Eagle Medium (DMEM) (Gibco, USA) supplemented with 100 U∙mL–1 penicillin and 10% fetal bovine serum (FBS; Biological Industries, Israel). After 1.5 h, to separate CMs from cardiac fibroblasts, 0.1 mmol∙L–1 5-bromo-2-deoxyuridine (BrdU) was added to inhibit fibroblast proliferation. The cells were then cultured for an additional 48 h post-plating. Treatments included scrambled negative control RNA (siNC), siNSUN2, siLARP1, siYBX1 synthesized by Guangzhou RiboBio Co., Ltd. (China), or adenovirus (AdV)-Null, AdV-NSUN2 in vitro (3.16 × 1010 plaque-forming units at a volume of 0.5 μL∙well–1 of six-well cell culture plates), or plasmid vector or LARP1. After being starved in serum-free DMEM (Gibco) for 6–8 h, CMs were treated with Ang II (1 μmol∙L–1) for 48 h.
2.9. RNA decay
CMs were cultured and treated with AdV-Null, AdV-NSUN2, siNC, siNSUN2, siYBX1, siGATA4, Ang II + siNC, or Ang II + siLARP1. Subsequently, cells were exposed to 5 μg∙mL–1 actinomycin D (MCE, USA) for 0, 3, or 6 h. The degradation rate of RNA was determined at the specified time points.
2.10. Quantitative real time-polymerase chain reaction (qRT-PCR)
qRT-PCR was performed using SYBR Green fluorescent dye (Toyobo, Japan) on the Applied Biosystems 7500 FAST Real-Time PCR System (Thermo Fisher, USA) to assay transcript levels of NSUN2, LARP1, its flag-tagged variant (LARP1-flag), GATA4, YBX1, atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), α-smooth muscle actin (α-SMA), collagen type I (Col1a1), and Collagen type III (Col3a1). Relative expression levels were normalized via the comparative threshold cycle (Ct) methodology (2−ΔΔCt), ensuring quantitative accuracy across experimental replicates.
2.11. m5C dot blot assay
RNA samples, normalized for concentration, were carefully spotted onto nylon membranes. The membranes were allowed to air-dry for 10 min before being crosslinked under ultraviolet irradiation. Following this, the spots were stained with methylene blue, and then blocked with 5% non-fat milk for 1 h. The membranes were incubated overnight with anti-m5C antibody (1:1000 in PBS, MABE1081; Sigma-Aldrich), following primary antibody binding, underwent a 90 min exposure to horseradish peroxidase (HRP)-labeled rabbit-specific immunoglobulin G (IgG) secondary antibody at ambient temperature to detect target proteins. Finally, visualization was achieved using enhanced chemiluminescence.
2.12. RNA immunoprecipitation (RIP) assay
The RIP protocol was executed using the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Merck, USA), adhering strictly to the supplier’s standardized guidelines for antibody incubation, magnetic bead separation, and RNA elution. Magnetic beads were pre-incubated with LARP1 antibody before adding the lysate supernatant for overnight incubation. After thorough washing of the beads, proteins were digested at 55 °C for 30 min, and the protease K digestion buffer was formulated with 117 μL of RIP wash buffer (1×), supplemented with 15 μL of 10% sodium dodecyl sulfate (SDS), and adjusted using 18 μL of Proteinase K (20 mg∙mL–1) to ensure efficient digestion while preserving RNA integrity. RNA was then extracted using the RIP wash buffer for subsequent qRT-PCR analysis.
2.13. Immunoblotting
Cell and tissue lysates were prepared using the lysis buffer (Beyotime, China) containing 1% protease inhibitor and 10% phosphatase inhibitor. Protein concentrations were quantified using a bicinchoninic acid (BCA) protein assay kit (Beyotime). Proteins were separated via sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and transferred onto nitrocellulose membranes. Densitometric analysis of Western blotting (WB) films was conducted using the Odyssey infrared imaging platform (LI-COR Biosciences, USA) to measure protein band intensities and gray value, which ensures quantitative precision in signal detection.
2.14. Methylated RNA immunoprecipitation (MeRIP)–qRT-PCR
A portion of CM-derived RNA was reserved as a technical replicate to normalize m5C modification levels, and the remaining RNA underwent immunoprecipitation (IP) using magnetic beads conjugated to anti-m5C antibodies. Parallel reactions included IgG-conjugated beads under identical incubation parameters (4 °C, rotation) to control for nonspecific binding. Post-IP, co-precipitated RNA from both m5C and IgG samples was purified and analyzed via qRT-PCR with locus-specific primers to quantify methylation-enriched transcripts.
2.15. Statistical analysis
All experiments were independently replicated three times, with outcomes expressed as mean ± standard error of the mean (SEM). Statistical evaluations were performed using GraphPad Prism 7.0 and SPSS software. Prior to analysis, data distribution was evaluated for normality. Parametric tests were applied to normally distributed data: the Student’s t-test compared two groups, while one-way analysis of variance (ANOVA) with appropriate post hoc comparisons assessed differences across over three groups. Non-parametric alternatives were used for skewed data: the Mann–Whitney U test for pairwise comparisons and the Kruskal–Walli’s test followed by Dunn’s multiple comparison test for three or more groups. Statistical significance was defined as P < 0.05.
3. Results
3.1. NSUN2 is upregulated in cardiac hypertrophy
To investigate the impact of RNA m5C modification on cardiac hypertrophy, we assessed the RNA m5C levels and examined the m5C methyltransferase expression in mouse hearts subjected to TAC. Dot blot assays revealed a significant elevation in RNA m5C levels in TAC-induced hypertrophic hearts (Figs. S1(a)–(g) in Appendix A). Among the RNA m5C methyltransferases, RNA sequencing (RNA-seq) analysis showed that
NSUN2 was notably upregulated in TAC-treated hearts compared to that in the Sham-operated controls, thus suggesting that NSUN2 may account for the increased RNA m5C level (
Fig. 1(a)). To further validate our hypothesis, we detected protein levels of NSUN2 in mouse hypertrophic hearts at various time points after TAC. The results demonstrated that TAC induced a progressive cardiac hypertrophy and functional impairment from three to ten weeks, which was concomitant with steadily increased mRNA levels of
NSUN2 at six and ten weeks, and immunofluorescence staining indicated that the colocation of NSUN2 with α-actinin showed a notable increase in hypertrophic hearts when compared to that in control hearts (
Figs. 1(b)–(d), Fig. S1(h) in Appendix A). Similarly, NSUN2 levels were significantly increased in an alternative murine model of cardiac hypertrophy induced by Ang II treatment, which were validated through both
in vivo and in
vitro experimental systems (
Figs. 1(e) and
(f), Fig. S1
(i) in Appendix A). Heart function was also assessed after continuous subcutaneous administration of Ang II for four weeks (Fig. S1(j) in Appendix A). To strengthen the clinical implication of our findings, we detected the NSUN2 expression in the cardiac tissues of patients diagnosed with HF. WB result showed the NSUN2 protein levels were strikingly elevated in failing human hearts compared to that in non-HF hearts (
Fig. 1(g), Table S1). Collectively, these results suggest that an increase in NSUN2-mediated m5C modification may play a role in the onset of pathological cardiac hypertrophy and HF.
3.2. Cardiomyocyte-specific knockout of NSUN2 attenuates pressure overload-induced cardiac hypertrophy and HF
To explore the function of NSUN2 in cardiac hypertrophy and HF, we generated conditional Cardiomyocyte (CM)-specific
NSUN2 knockout (
NSUN2-cKO) mice by crossing conditional knockout allele of
NSUN2 (
NSUN2flox/flox) mice with
Myh6-Cre mice. The
NSUN2-cKO mice were viable, fertile, and phenotypically normal. No statistically significant variations in body weight or heart rate were detected between NSUN2-cKO and
NSUN2flox/flox mice at ten weeks of age (Fig. S2(a) in Appendix A). WB and qRT-PCR demonstrated significant reductions in
NSUN2 mRNA and protein levels in the hearts of
NSUN2-cKO mice compared to those in
NSUN2flox/flox mice, as demonstrated by WB and qRT-PCR (Figs. S2(b)–(d) in Appendix A). To explore the consequences of NSUN2 deficiency on cardiac hypertrophy, both
NSUN2-cKO and
NSUN2flox/flox mice were subjected to Sham or TAC surgery. Echocardiographic analyses revealed no statistical differences in terms of cardiac functional and structural features between
NSUN2-cKO and
NSUN2flox/flox mice after the Sham operation (Figs. 2(a)–(c), Figs. S2(e) and (f) in Appendix A). However, after ten weeks of TAC, cardiac function was impaired in NSUN2
flox/flox mice, whereas
NSUN2-cKO mice showed improved cardiac function, as evidenced by increased ejection fraction (EF) and fractional shortening (FS) (
Figs. 2(a) and
(b)). The
NSUN2-cKO mice also exhibited less cardiac dilation after TAC than
NSUN2flox/flox mice, as displayed reduced LVID;d and LVID;s (Fig. S2(f)). In addition to functional improvements,
NSUN2-cKO mice also showed a remarkable alleviation of cardiac hypertrophy compared to
NSUN2flox/flox controls after TAC, which was reflected by reductions in LVPW;d, the ratios of heart weight to body weight (HW/BW) and heart weight to tibial length (HW/TL), and a decrease in CM cross-sectional area (CSA) (Figs. 2(c)–(e), Fig. S2(e)). Next, we measured the mRNA expression of two hypertrophic genes and found that
ANP and
BNP were significantly upregulated in TAC-treated
NSUN2flox/flox mice, which were attenuated in
NSUN2-cKO mice (
Fig. 2(f)). Furthermore, Masson’s trichrome staining showed that
NSUN2-cKO mice exhibited less cardiac fibrosis after TAC than NSUN2flox/flox littermate controls (
Figs. 2(g) and
(h)). Meanwhile, knockout of
NSUN2 downregulated the levels of fibrosis marker genes (Fig. 2(i), Fig. S2(g) in Appendix A).
To further substantiate our findings, we employed an alternative, well-established mouse model of cardiac hypertrophy induced by continuous subcutaneous infusion of Ang II. Similar to the TAC model, Ang II-induced cardiac hypertrophy in mice also increased the mRNA and protein expression of NSUN2, whereas silencing of NSUN2 effectively reduced the expression of NSUN2 in the heart (Figs. S3(a) and (b) in Appendix A). Interestingly, the NSUN2-cKO and NSUN2flox/flox mice subjected to Ang II infusion for four weeks exhibited a significant increase in both systolic and diastolic blood pressure compared to the saline-treated control, as determined by telemetry. Both genotypes exhibited a similar level of blood pressure after Ang II infusion (Fig. S3(c) in Appendix A). As anticipated, a pronounced cardiac hypertrophy was induced by Ang II infusion in NSUN2flox/flox mice, as indicated by the remarkable increases in HW/BW, HW/TL, LVID;d, LVID;s, LVPW;d, and CSA, whereas these effects were largely abrogated in NSUN2-cKO mice (Figs. S3(d)–(g) in Appendix A). Meanwhile, the increased ANP and BNP mRNA levels induced by Ang II were suppressed in NSUN2-cKO mice compared to those in littermate controls (Fig. S3(h) in Appendix A). Moreover, NSUN2-cKO mice showed notably improved cardiac function and reduced cardiac dilatation upon Ang II administration compared with those in NSUN2flox/flox mice (Figs. S3(i) and (j) in Appendix A). Masson’s trichrome staining showed that NSUN2-cKO mice exhibited less cardiac fibrosis after Ang II treatment than that in NSUN2flox/flox mice. In line with these findings, the Ang II-induced increase in the mRNA expression of fibrotic markers was also reduced by NSUN2 silencing (Figs. S3(k)–(m) in Appendix A). Collectively, these results indicate that NSUN2 is essential for triggering cardiac hypertrophy and HF under conditions of pressure overload.
3.3. Cardiac specific overexpression of NSUN2 causes cardiac hypertrophy
These results presented above prompted us to further investigate whether NSUN2 overexpression affects cardiac hypertrophy. For this purpose, we engineered AAV9 vectors carrying
NSUN2 under the control of a troponin T promoter (AAV9-NSUN2) to overexpress NSUN2 in mouse heart tissue via tail vein injection. As shown in Figs. S4(a) and (b) in Appendix A, NSUN2 expression was significantly increased after treatment with AAV9-
NSUN2 for eight weeks, reaching a level similar to that in TAC-treated mouse hearts. Moreover, overexpression of NSUN2 at the cellular level or use of Ang II to induce CM hypertrophy significantly increased the level of m5C methylation modification (Fig. S4(c) in Appendix A). Echocardiographic analysis showed that cardiac-overexpression of NSUN2 induced cardiac dysfunction and dilatation as indicated by decreased EF and FS, and increased LVID;d and LVID;s (
Figs. 3(a) and
(b), Fig. S4(d) in Appendix A). Notably, cardiac-overexpression of NSUN2 led to hypertrophic phenotypes, as indicated by increases in HW/BW, HW/TL, and LVPW;d (
Fig. 3(c), Fig. S4(e) in Appendix A). WGA staining further verified that the CSA was robustly increased in NSUN2-overexprerssing hearts (
Figs. 3(d) and
(e)). Similarly, the upregulation of NSUN2 led to a marked elevation in the mRNA expressions of
ANP and
BNP (
Fig. 3(f)). Furthermore, NSUN2 overexpression resulted in remarkable increases in cardiac fibrosis and in mRNA levels of
Col1a1,
Col3a1, and
α-SMA (
Figs. 3(g)–(i)). These findings indicate that NSUN2 expression alone can trigger cardiac hypertrophy and remodeling. To corroborate those
in vivo results, we used AdV-
NSUN2 to infect CMs and induce NSUN2 overexpression
in vitro. In the
in vivo results, overexpression of NSUN2 was sufficient to induce CM hypertrophy, as demonstrated by enhanced cell surface area and ANP and
BNP mRNA levels (Figs. S5(a)–(c) in Appendix A). Next, we performed a loss-of-function to knock down endogenous
NSUN2 by transfecting with
NSUN2 small interfering RNA (siNSUN2) or siNC construct into neonatal mouse cardiomyocytes (NMCMs), followed by the establishment of CM hypertrophy induced by Ang II. The knockdown efficacy of
NSUN2 expression was confirmed, and this downregulation resulted in a significant decrease in the m5C modification levels (Figs. S5
(d)–(f) in Appendix A). As anticipated, Ang II treatment increased the cell surface area as determined by immunofluorescence staining, which was markedly mitigated by NSUN2 knockdown in NMCMs (Fig. S5(g) in Appendix A). Silencing of
NSUN2 repressed Ang II-induced increases in
ANP and
BNP mRNA levels (Fig. S5(h) in Appendix A).
3.4. LARP1 is the target of NSUN2
This study aims to uncover the molecular pathways and cellular processes through which NSUN2 drives pathological cardiac hypertrophy. To this end, we conducted m5C MeRIP sequencing (MeRIP-seq) in hypertrophic mouse hearts (Fig. S6(a) in Appendix A). Consistent with published report [
8], m5C peaks preferred the GC-enriched motifs in the mouse myocardium (Fig. S6
(b) in Appendix A). Furthermore, the majority of m5C peaks were found to be enriched in the coding sequence (CDS), start codon (startC), or stop codon (stopC), with a smaller number of m5C peaks being enriched in the 5′ untranslated region (UTR) or 3' UTR (Figs. 4(a) and
(b)). Next, we used MeRIP-seq and RNA-seq results to correlate the level of m5C peaks with their genes expression, and identified 507 upregulated genes with 1021 hypermethylated m5C peaks (termed as hyper-up) in TAC hearts relative to that in Sham-operated controls (
Fig. 4(c)). Gene Ontology (GO) functional analysis showed that the differentially expressed m5C methylated genes were mainly involved in cell growth and organism growth (Fig. S6
(c) in Appendix A). The Kyoto Encyclopedia of Genes and Genomes (KEGG) functional analysis revealed an enrichment of differentially expressed m5C methylated genes in multiple pathways, such as the Wingless/integration-1 (Wnt) signaling pathway, and hypertrophic cardiomyopathy disease (Fig. 4(d)). Given that the increased global m5C levels in TAC-induced hypertrophic hearts, we predicted that the upregulated genes are potential targets of NSUN2. A bioinformatic intersection analysis between the upregulated gene and differentially expressed genes from the database of human HF samples (GSE21610), identified 42 genes (Fig. S6
(d) in Appendix A). To identify the downstream targets of NSUN2, we focused on genes enriched in the Wnt signaling pathway, which along with the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB/AKT) pathway has been implicated in the pathogenesis of cardiac hypertrophy [17,
18]. We performed an m5C-MeRIP–qPCR assay to measure the m5C modification levels of some genes with relatively high abundance in the heart tissue, including
LARP1, human immunodeficiency virus type I enhancer binding protein 1 (HIVEP1), secreted frizzled-related protein 1 (
SFRP1), proline arginine-rich end leucine-rich repeat (PRELP), growth factor receptor bound protein 10 (GRB10), and microtubule associated monooxygenase, calponin, and LIM domain containing 2 (MICAL2). The results showed that the m5C levels of
LARP1 were markedly increased after
NSUN2 overexpression (
Fig. 4(e)). Moreover, the level of LARP1 was elevated in both TAC- and Ang II-treated mice, and was attenuated by
NSUN2 knockdown (Figs. S6(e) and (f) in Appendix A). Similarly,
LARP1 m5C levels increased in
NSUN2flox/flox mice after TAC, whereas this effect was significantly abolished in
NSUN2-cKO hearts (
Fig. 4(f)). Integrative Genomics Viewer (IGV)-based interrogation revealed elevated m5C modification prevalence within the CDS of
LARP1 transcripts in hypertrophic myocardium, concomitant with a significant upregulation in
LARP1 mRNA levels compared to Sham-operated controls (
Fig. 4(g)). Given that the m5C modification regulates the transcript stability [
19], RNA decay assays conducted on actinomycin D-treated NMCMs assessed the role of NSUN2 modulating
LARP1 transcript stability. The results showed that NSUN2 overexpression attenuated, whereas
NSUN2 knockdown accelerated
LARP1 mRNA degradation (Fig. 4(h)). Temporal profiling of LARP1 protein dynamics in TAC mouse myocardium revealed a biphasic upregulation pattern, characterized by significant elevation at the third week, remained stable, and increased significantly at the tenth week (Fig. S6(g) in Appendix A). Furthermore, LARP1 protein levels were markedly elevated in TAC-treated
NSUN2flox/flox mice; however, this effect was reversed in
NSUN2-cKO hearts (
Fig. 4(i)). Consistently,
NSUN2 knockdown reduced Ang II-induced upregulation of LARP1 protein expression in NMCMs (Fig. S6(h) in Appendix A). In contrast, NSUN2 overexpression markedly elevated LARP1 protein expression (Fig. S6
(i) in Appendix A). These data demonstrate that LARP1 functions as a direct downstream substrate of NSUN2-catalyzed m5C RNA modification, underscoring its role as a regulatory node in epitranscriptomic signaling. 3.5. Knockdown of LARP1 abrogates TAC-induced cardiac hypertrophy
LARP1 is an RNA-binding protein that influences RNA stability and translation [20,
21] and is involved in the Wnt and PI3K/AKT signaling pathway [[22],
[23],
[24]], suggesting that LARP1 may be related to cardiac hypertrophy. To address this question, we performed an
in vitro study to knockdown
LARP1 in NMCMs using siRNA-mediated transfection. Three
LARP1-specific siRNAs were used to silence
LARP1 in NMCMs, and we found that si
LARP1#1 showed the highest knockdown efficiency (Figs. S7(a) and (b) in Appendix A). Accordingly, this siRNA was used in subsequent
LARP1 knockdown studies. Knockdown of
LARP1 attenuated Ang II-induced upregulation of
LARP1,
ANP, and
BNP mRNA and proteins levels, and concurrently reduced CM hypertrophy (Figs. S7(c)–(e) in Appendix A). To further corroborate the function of LARP1 in cardiac hypertrophy
in vivo, the AAV9-sh
LARP1 construct was delivered to mice via tail vein injection to silence endogenous LARP1 in the heart before TAC surgery. As shown in Figs. S8(a) and (b) in Appendix A, the silencing efficiency of
LARP1 at both the mRNA and protein levels in the mouse heart was verified by WB and qRT-PCR. Moreover, administration of AAV9-sh
LARP1 also attenuated the TAC-induced elevation of
LARP1 mRNA and protein levels compared to those in the AAV9-shNC-treated group. Echocardiography revealed no notable disparities in the morphological and functional parameters between mice treated with AAV9-shNC and AAV9-shLARP1 in the absence of TAC operation. However, TAC induced cardiac dysfunction as reflected by decreases in EF and FS, and cardiac dilatation was alleviated by LARP1 knockdown (
Figs. 5(a) and
(b)). In line with the
in vitro results,
LARP1 silencing markedly attenuated cardiac hypertrophic growth as displayed by reduced LVPW;d, LVID;s, LVID;d, HW/BW, and HW/TL (
Fig. 5(c), Figs. S8(c) and (d) in Appendix A). Similarly,
LARP1 knockdown reduced the CSA of CMs and the expression of
ANP and
BNP upon TAC (
Figs. 5(d)–(f)). Furthermore,
LARP1 silencing diminished cardiac fibrosis in TAC-treated mice (
Figs. 5(g)–(i)). These results suggested that LARP1 contributes to cardiac hypertrophy following pressure overload.
To determine whether LARP1 regulates NSUN2-driven cardiac hypertrophy, we conducted rescue experiments. Results showed that the NSUN2-induced upregulation of LARP1 in mouse heart was abolished by treatment with AAV9-shLARP1 (
Fig. 6(a)). Restoration of LARP1 expression alleviated NSUN2 overexpression-induced cardiac dysfunction and cardiac dilatation (
Figs. 6(b) and
(c), Figs. S9(a) and (b) in Appendix A). Furthermore, knockdown of
LARP1 cancelled out the pro-hypertrophic response induced by NSUN2 overexpression, as shown by the decrease in LVPW;d, HW/BW, and HW/TL (Figs. S9(a) and (b)). WGA staining confirmed that the NSUN2-induced increased in CSA was attenuated by
LARP1 silencing (
Figs. 6(d) and
(e)). In agreement with the
in vivo results, the NSUN2 overexpression-induced upregulation of cardiac hypertrophic genes (
ANP,
BNP) was attenuated by the restoration of LARP1 in NMCMs (Fig. S9(c) in Appendix A). Furthermore, Masson’s trichrome staining showed that the NSUN2-induced increase in fibrosis level was attenuated by
LARP1 silencing (
Figs. 6(f) and
(g)). These data demonstrated that LARP1 is a direct downstream mediator of NSUN2 activity during cardiac hypertrophy.
3.6. LARP1 regulates GATA4 during cardiac hypertrophy
To clarify the mechanism underlying the pro-hypertrophic effect of LARP1, we anticipated the potential downstream targets of LARP1 based on the RNA binding protein database (catRAPID omics v2.0). GO biological function analysis showed that these genes predicted to interact with LARP1 were related to signal transduction, positive regulation of gene expression, and heart development (Fig. S10(a) in Appendix A). Subsequently, we conducted an additional screening of genes associated with cardiac hypertrophy from catRAPID, which was further validated through bioinformatics overlap analysis with four additional databases (CTD, GeneCards, GSE36961, and GSE2014); we identified
GATA4 and
EGR1 as potential targets of LARP1 (Fig. S10(b) in Appendix A). To validate our hypothesis, we performed a RIP assay to validate their interaction with LARP1; results showed that LARP1 bound to
GATA4 and
EGR1 mRNA, respectively (
Fig. 7(a)). More importantly,
LARP1 knockdown markedly reduced
GATA4 mRNA and protein levels but not
EGR1 mRNA levels, in both TAC-treated hearts and Ang II-treated CMs (
Fig. 7(b), Figs. S10
(c) and (d) in Appendix A). Consistently, the protein expression level of GATA4 in the hearts of TAC mice increased at three weeks of age, and there was a significant increase in the protein expression at six and ten weeks (Fig. S10(e) in Appendix A). Given that GATA4 is a critical pro-hypertrophic transcriptional factor, we focused our attention on it [
25]. We then investigated the effect of LARP1 on
GATA4 stability by conducting an mRNA decay assay in CMs after
LARP1 silencing. As expected,
LARP1 knockdown accelerated the velocity of
GATA4 mRNA degradation in CMs treated with Ang II (Fig. 7(c)). Consistently, knockdown of
LARP1 notably reduced the mRNA and protein levels of
GATA4 with or without Ang II and TAC (
Fig.7(d), Figs. S10(c) and (d)). We also performed MeRIP–qPCR to determine whether NSUN2 regulates
GATA4 methylation. The result showed that overexpression of NSUN2 failed to increase the m5C levels of
GATA4 mRNA compared to those in the null-treated group, thus indicating that NSUN2 does not directly act on
GATA4 mRNA, but indirectly regulates
GATA4 stability through LARP1 (Fig. S10
(f) in Appendix A). Furthermore, catRAPID predicted that the RNA binding domain of GATA4 and LARP1 is located at amino acids 382–438 of LARP1 (Fig. S10(g) in Appendix A). These results suggested that LARP1 directly binds to
GATA4 mRNA to prevent its degradation, thereby contributing to cardiac hypertrophy.
To investigate whether GATA4 similarly serves as a downstream effector mediated by NSUN2 in the regulation of cardiac hypertrophy, through qRT-PCR and WB analyses, we quantified the expression of the target molecule across both transcriptomic and proteomic levels. Our results showed that TAC- and Ang II-induced cardiac hypertrophy significantly increased GATA4 expression, whereas silencing of NSUN2 both
in vivo and
in vitro markedly reduced
GATA4 mRNA and protein levels. Furthermore, Ang II-induced upregulation of GATA4 was reversed by
NSUN2 knockdown in NMCMs (
Fig. 7(e)). In addition, knockout of
NSUN2 markedly decreased
GATA4 mRNA and protein levels in the heart with and without TAC surgery (
Figs. 7(f) and
(g)). NSUN2 overexpression elevated the levels of GATA4, which were reversed by
LARP1 knockdown in the mouse myocardium (
Fig. 7(h), Fig. S11
(a) in Appendix A). In addition, downregulation of GATA4 expression after siGATA4 treatment repressed the NSUN2-induced upregulation of
ANP and
BNP mRNA expression (Figs. S11(b) and (c) in Appendix A). These data suggested that the LARP1–GATA4 axis mediates the pro-hypertrophic effects of NSUN2 on cardiac hypertrophy.
3.7. NSUN2 regulates LARP1 stability in an YBX1-m5C-dependent manner
These results suggested that NSUN2 enhanced
LARP1 stability by increasing its mRNA m5C modification. Since m5C modifications regulate mRNA stability, translation, and nuclear export [
7], we next examined the levels of m5C readers in TAC-treated mice. The results revealed that YBX1, but not ALYREF, was markedly upregulated in the hypertrophic mouse hearts (Figs. S12
(a) and (b) in Appendix A). Given that the established function of YBX1 in the maintenance of RNA stability [26], we hypothesized that YBX1 regulates
LARP1 mRNA stabilization. To address this issue, we conducted an RNA decay assay to assess the mRNA expression levels of LARP1 in NMCMs following
YBX1 knockdown. As anticipated, knockdown of
YBX1 accelerated the degradation of
LARP1 mRNA following actinomycin D treatment, suggesting that YBX1 is capable of increasing LARP1 expression by maintaining its mRNA stability (
Fig. 8(a)). Additional results supported the finding that the knockdown of
YBX1 significantly decreased the levels of LARP1 in both the absence and presence of Ang II stimulation in NMCMs (
Figs. 8(b)–(d), Fig. S12(c) in Appendix A). Based on the bioinformatics analysis showing that NSUN2-methylated m5C modification is mainly located in the CDS of
LARP1 mRNA, we constructed
LARP1 wild-type (
LARP1-WT) and the m5C sites mutant
LARP1 CDS (
LARP1-Mut) plasmids (
Fig. 8(e)) [
27,
28]. Transfection with
LARP1-WT or
LARP1-Mut plasmid led to an equal amount of increased protein level of LARP1 in NMCMs (Fig. S12(d) in Appendix A). MeRIP–qPCR result showed that NSUN2 overexpression increased the m5C level of
LARP1 mRNA in the
LARP1-WT group, but not in the
LARP1-Mut group, suggesting that these sites are indispensable for
LARP1 m5C modification (
Fig. 8(f)). Therefore, we performed an RNA decay assay to confirm whether
LARP1 m5C modification at these sites was vital for maintaining its mRNA stability. Our findings revealed that NSUN2 overexpression significantly stabilized
LARP1 mRNA in wild-type cells compared to that in knock-in mutants, which indicates enhanced transcript preservation under enforced NSUN2 activity (
Fig. 8(g)).
Next, we investigated whether YBX1 was involved in CM hypertrophy. To assess this, we conducted a loss-of-function experiment to silence the
YBX1 expression in NMCMs before Ang II infusion. As illustrated in Fig. S12(e) in Appendix A, we observed that the knockdown of
YBX1 mitigated the elevation in the mRNA levels of hypertrophic genes caused by Ang II. This result suggests that YBX1 can promote CMs hypertrophy upon pathological stimulation, which is supported by a recent study [
29]. Moreover,
YBX1 knockdown reduced
GATA4 mRNA and protein expressions in NMCMs, whether Ang II treatment was applied or not (Figs. S12(f) and (g) in Appendix A). To verify that NSUN2-induced CM hypertrophy is mediated by YBX1, we silenced
YBX1 in NSUN2-overexpressing CMs. The results showed that silencing of
YBX1 notably reduced the increase in
ANP and
BNP levels induced by NSUN2 overexpression. Meanwhile, NSUN2 caused a prominent increase in
LARP1 and
GATA4 mRNA levels, whereas these effects were markedly abrogated by silencing of
YBX1. Furthermore,
GATA4 knockdown rescued the hypertrophic phenotype induced by NSUN2, as indicated by decreased
ANP and
BNP mRNA levels (
Fig. 8(h)). Taken together, NSUN2 increased LARP1 expression in a YBX1/m5C-dependent manner.
4. Discussion
HF is closely associated with cardiac hypertrophy and promotes its initiation and progression. In patients with HF, we detected an increased expression of NSUN2, a methyltransferase involved in m5C modification, within the myocardium. NSUN2 triggers cardiac hypertrophy by enhancing LARP1 stability in a YBX1/m5C-dependent manner, thereby preventing the degradation of GATA4. Deletion of NSUN2 or LARP1 exhibited beneficial effects on cardiac remodeling by repressing GATA4 expression. These findings may provide novel knowledge on RNA modification in the pathogenesis of cardiac hypertrophy.
Epigenetic modifications, which are pervasive in both physiological and pathological processes of the human body, are mainly categorized into three types: histone, DNA, and RNA modifications. The functional significance of RNA modifications like m6A and m1A has been widely documented [30]. They can exert their effects by binding to the 5′ UTR, 3′ UTR, and CDS regions of target RNAs to regulate RNA stability and translation [31,
32]. In this study, we found that NSUN2 binds to the CDS region of
LARP1 to regulate its RNA stability. Moreover, as an important m5C “writer,” by modulating the efficiency of
ICAM-1 mRNA translation, NSUN2 deficiency could protect endothelial cells from inflammatory damage [16], and reduce the m5C methylation of tRNA and the subsequent production of overall protein synthesis, thereby reducing liver injury [33]. The current research findings suggest that, while
NSUN2 silencing reduces global protein expression, it may also attenuate the expression of hypertrophy-related genes under cardiac injury conditions. Based on the RNA-seq and epigenomics data, we explored the potential downstream targets of NSUN2, and screened for targets involved in pathways closely related to cardiac hypertrophy, such as cardiac metabolism, cell and tissue growth, inflammation, immunity, and apoptosis, using GO enrichment analysis. Through experimental validation, we identified LARP1, an RNA-binding protein associated with cell growth, as the binding partner of NSUN2. LARPs are RNA-binding proteins that feature the La module, consisting of LAM and RRM motifs, as well as unique regions specific to LARP1–7 members [34]. LARP7 is a major regulator of HF [
35]. Additionally, RNA-binding proteins are known to regulate RNA stability, translation, and
NF-κB mRNA stabilization during T cell activation [36]. As another La-related protein that may participate in the regulation of cardiac function,
LARP1 gene contains two functional domains, the La module and DM15 for mRNA cap-binding. Among them, either the DM15 domain or LAM domain of
LARP1 can directly bind to poly(A) and poly(A) binding protein, cytoplasmic 1 (PABPC1) [
37]. However, only the C-terminal DM15 domain of LARP1 has a specific binding motif for 5′ terminal oligopyrimidine (
TOP) mRNA [
38]. Moreover, LARP1 can competitively bind to the 5′-ends of mRNAs of eukaryotic initiation factor 4F (eIF4F) gene [
20], and it reportedly affects RNA stability through binding to the 3′ UTR of myelocytomatosis (MYC) oncogene, B cell leukemia/lymphoma 2 (BCL2), and BCL2-interacting killer (
BIK), thus impacting tumor progression [
21,
39,
40]. It is also thought to promote lung cancer cell growth by sequestering miR-330-5p [
41]. However, the effect of LARP1 on cardiac disease has not been elucidated. Among LARPs family members, LARP1 has a La module for binding to the PABPC1 and DM15 domains to influence TOP mRNAs translation [
37]. According to recent research, double mutations of the two residues Q333A/F348A and Y336A/F348A in the LaM domain in human cells abolished the affinity of LARP1 binding [42]. Subsequently, by using website prediction (catRAPID omics v2.0), we found that the functional domain for RNA binding of LARP1 to GATA4 is at amino acids 382–438 (Fig. S10(g) in Appendix A), which aligns with the findings of Kozlov et al. [
42]. Additionally, the residues 669–1019 interact with
TOP mRNA in a mechanistic target of rapamycin complex 1 (mTORC1)-dependent manner [37]. The existence of numerous RNAs containing TOP structural domains has been widely reported, while in enrichment analysis of some of the clearly reported genes which are potentially downstream of LARP1, we did not find significant enrichment in our current study. Moreover, by analyzing and validating the downstream of GO pathways (enrichment in energy metabolism, mitochondrial function, and cell proliferation and growth, which are highly associated with cardiac hypertrophy), we found a high binding capacity between GATA4 and LARP1. Although LARP1 regulates RNA stability and translation, its roles in many pathological processes remain unexplored. Overwhelming evidence supports that LARP1 can regulate cancer cell growth and proliferation and reduce apoptosis in various cancers [
21,
39,
40]. In the current study, we found in the present study that LARP1 affected the mRNA stability of
GATA4 mRNA.
The zinc finger-containing transcription factor GATA4 serves as a central mediator in pathological cardiac remodeling by orchestrating the transcriptional activation of embryonic cardiac genes, which is critical for the development of maladaptive hypertrophy [25]. Dysregulations of various post-translational modifications of GATA4, which are essential for altering its transcriptional activity as a pro-hypertrophic factor, are observed during cardiac hypertrophy and HF. According to studies on GATA4, when acetylated or phosphorylated, its transcriptional activity is enhanced to produce hypertrophic genes, whereas when deacetylated, it exhibits the opposite effect. Conversely, other studies have demonstrated that the phosphorylation of GATA4 enhanced while its methylation decreased
GATA4 transcriptional activity during cardiac hypertrophy [
43]. Some studies have indicated that certain proteins, as coactivators, interact with GATA4 in the nucleus to activate its transcriptional activity [
44]. Additionally, miRNAs have been reported to repress GATA4 protein expression at the post-transcriptional level [45], and the RNA-binding protein, FUS, influences the splicing pattern and subcellular distribution of GATA4 by binding to its mRNA [
46]. In contrast, our study showed that LARP1 directly bound to
GATA4 mRNA, and thereby preventing its degradation in NMCMs. Overexpression of NSUN2 or LARP1 exacerbated TAC- or Ang II-induced increases in
GATA4 mRNA and protein levels in both myocardial tissue and NMCMs, whereas genetic ablation of these factors suppressed hypertrophic signaling. Notably, siRNA-mediated GATA4 knockdown reversed NSUN2/LARP1-driven CM enlargement and reduced
ANP/
BNP transcript levels in NMCMs under hypertrophic stress. Collectively, these findings establish the NSUN2–LARP1/GATA4 axis as a central regulatory pathway governing pathological cardiac hypertrophy.
YBX1, a pleiotropic nucleic acid-binding protein, orchestrates critical cellular functions spanning DNA replication, transcriptional regulation, mRNA stability maintenance, and translational control through its ability to bind both DNA and RNA. Evidence suggests YBX1 serves as a central regulator of diverse pathophysiological processes, which encompass canonical pathways such as cellular proliferation, differentiation programs, apoptotic execution, and responses to oxidative or genotoxic stress [47,
48]. Recent studies have shown that YBX1 acts as an m5C reader to maintain mRNA stability and enhances translational efficiency by recruiting the RNA-binding protein ELAVL1 [7]. Another m5C reader, ALYREF, regulates nuclear-cytoplasmic shuttling of RNA [
8]. Moreover,
YBX1 knockdown inhibits CM proliferation [
49]. A previous study showed that YBX1 protein levels were increased in both TAC-treated mouse hearts and phenylephrine-treated CMs, and this elevation promoted protein synthesis and cellular growth by binding to and upregulating eukaryotic elongation factor 2 [29]. They further showed that depletion of
YBX1 protected against cardiac hypertrophy and dysfunction under pressure. In line with this study, our data showed that YBX1, but not ALYREF, was robustly increased in the hypertrophic heart tissues after TAC. More importantly,
YBX1 knockdown alleviated either Ang II- or NSUN2-induced CM hypertrophy by repressing LARP1 and GATA4 expression. A mechanistic study demonstrated that YBX1 was unable to recognize the mutated m5C-modified sites of
LARP1 transcript, thereby losing its ability to maintain the stability of
LARP1 mRNA.
Our study has some limitations, since YBX1 affects mRNA translation after recognizing m5C-methylated transcripts, our current results failed to uncover whether the increased m5C levels in LARP1 mRNA can also alter its translation capability. Further studies are required to confirm this hypothesis. Although we demonstrated that the LARP1 binds to GATA4 mRNA to prevent its degradation, our data did not determine the specific domain responsible for binding to GATA4 mRNA. Additionally, the mechanism underlying the increased stability of GATA4 mRNA induced by LARP1 needs to be elucidated.
5. Conclusions
Taken together, our results showed that the m5C methyltransferase NSUN2 was elevated in the myocardium of patients with HF and mice with hypertrophic heart tissue in response to pressure overload and Ang II. NSUN2 enhances the stability of LARP1 in a YBX1/m5C-dependent manner and prevents the degradation of GATA4, which triggers cardiac hypertrophy. Loss of NSUN2 or LARP1 has a beneficial effect on cardiac remodeling by inhibiting the expression of GATA4. The current study advances our understanding of molecular mechanisms underpinning cardiac hypertrophy by elucidating novel pathways implicated in its pathological progression.
CRediT authorship contribution statement
Yingqi Liu: Writing – review & editing, Writing – original draft. Fan Wu: Validation. Kuiwu Liu: Conceptualization. Shuting Yu: Project administration. Changhao Wang: Formal analysis. Xin Li: Formal analysis. Zhiyong Sun: Formal analysis. Wanhong Li: Validation. Yi Zhang: Validation. Tiantian Ju: Formal analysis. Qian Liu: Formal analysis. Min Huang: Investigation. Zhongting Mei: Investigation. Zhezhe Qu: Investigation. Meixi Lu: Conceptualization. Xiaochen Pang: Conceptualization. Yingqiong Jia: Conceptualization. Jianhao Jiang: Conceptualization. Shunkang Dou: Conceptualization. Na Li: Validation. Yaozhi Zhang: Validation. Ying Li: Validation. Chuanhao Huang: Validation. Yuechao Dong: Validation. Baofeng Yang: Supervision. Weijie Du: Writing – original draft, Supervision, Writing – review & editing, Validation, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by grants from the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0537909), the National Natural Science Foundation of China (82273928, 82473919, 82330011, and 82161148007), the Distinguished Young Scholars of Natural Science Foundation of Heilongjiang Province (JQ2024H002), the Chunyan Programme of Heilongjiang Province (CYQN2403), and the CAMS Innovation Fund for Medical Sciences (CIFMS, 2020-I2M-5-003).
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2025.05.016.
PDF
(4885KB)
