GRSF1 Protects Against Heart Failure by Maintaining BCAA Homeostasis

Abstract

BACKGROUND:

Imbalances in cardiac branched-chain amino acid (BCAA) metabolism and mitochondrial homeostasis are implicated in the onset and development of heart failure. However, the mechanisms triggering the downregulation of cardiac BCAA metabolism in heart failure remain unclear. Here, we identify a novel role of the RNA-binding protein GRSF1 (guanine-rich RNA sequence binding factor 1) in post-transcriptionally regulating cell-intrinsic BCAA metabolic pathways, ultimately contributing to the pathogenesis of heart failure.

METHODS:

We examined GRSF1 expression in the heart tissues of patients with dilated cardiomyopathy and generated mice with cardiomyocyte-specific deletion or overexpression of GRSF1 in vivo to investigate its role in heart failure. The effect of GRSF1 on BCAA homeostasis was assessed through untargeted and targeted metabolomics and mitochondrial function analysis. To elucidate the mechanisms underlying GRSF1-mediated metabolic regulation, we employed mice with cardiomyocyte-specific deletion of BCKDHB (branched-chain keto acid dehydrogenase E1 subunit β) and mice with cardiomyocyte-specific expression of GRSF1 lacking a quasi-RNA recognition motif.

RESULTS:

GRSF1 expression was significantly decreased in the hearts of patients with heart failure and failing murine hearts. Cardiomyocyte-specific GRSF1 deletion resulted in cardiac dysfunction, spontaneous progression to dilated cardiomyopathy, and heart failure, accompanied by increased cardiac hypertrophy and fibrosis. Conversely, GRSF1 overexpression attenuated cardiac remodeling and heart failure induced by transverse aortic constriction. Mechanistically, GRSF1 maintained BCAA homeostasis and mitochondrial function by directly interacting with the G-tracts in the coding region of BCKDHB mRNA through a quasi-RNA recognition motif to promote the stability of BCKDHB mRNA at the post-transcriptional level, thereby increasing its protein expression. Functional recovery mediated by GRSF1 overexpression in cardiomyocytes was partially blocked upon cardiac-specific deletion of BCKDHB.

CONCLUSIONS:

Our study identified GRSF1 as a cell-intrinsic metabolic checkpoint that maintains cardiac BCAA homeostasis by regulating BCKDHB mRNA turnover. Targeting GRSF1 may offer therapeutic benefits for heart failure and other cardiometabolic diseases requiring BCAA manipulation.

Supplementary Material

cir-153-736-s001.pdf ^{(2.3MB, pdf)}

cir-153-736-s002.pdf ^{(15.1MB, pdf)}

cir-153-736-s003.pdf ^{(189.4KB, pdf)}

cir-153-736-s004.pdf ^{(269.4KB, pdf)}

cir-153-736-s005.docx ^{(16.1KB, docx)}

Clinical Perspective.

What Is New?

• GRSF1 (guanine-rich RNA sequence binding factor 1) expression is decreased in hearts from patients and mice with heart failure.

• Cardiac GRSF1 depletion results in dilated cardiomyopathy and heart failure, whereas GRSF1 overexpression in cardiomyocytes protects against metabolic and structural dysfunction in a heart failure model.

• GRSF1 maintains branched-chain amino acid homeostasis and mitochondrial function by stabilizing BCKDHB (branched-chain keto acid dehydrogenase E1 subunit β) mRNA at the post-transcriptional level.

What Are the Clinical Implications?

• GRSF1 represents a potential therapeutic target for modulating cardiac metabolism in heart failure.

• Restoring GRSF1 expression could stabilize BCKDHB mRNA, mitigate metabolic dysregulation, and improve cardiac outcomes in patients with heart failure.

• Therapeutic strategies focused on branched-chain amino acid metabolic homeostasis may address an unmet need in cardiometabolic disease management.

Heart failure, the end stage of various cardiovascular diseases, is the leading cause of hospitalization and mortality worldwide.¹ The etiology of heart failure is multifaceted, with metabolic disorders, particularly imbalances in mitochondrial metabolic homeostasis, potentially playing a significant role.^2,3 The branched-chain amino acids (BCAAs) isoleucine, leucine, and valine exhibit disrupted homeostasis in patients with heart failure.⁴ Beyond serving as building blocks for protein synthesis, BCAAs yield an array of metabolic intermediates, some of which have unique signaling properties that influence heart failure progression.⁵ BCAAs undergo reversible transamination, mediated by mitochondrial BCAA aminotransferase, to yield branched-chain α-keto acids (BCKAs). BCKAs are subsequently degraded by the BCKA dehydrogenase (BCKDH) complex, a rate-limiting enzyme complex in this process.⁶ Maintaining BCKDH complex activity and expression is, therefore, critical for restoring BCAA metabolic homeostasis in heart failure. However, the precise upstream control points of dysfunctional BCAA metabolism, particularly in heart failure, are not well understood.

Post-transcriptional regulation of mRNA is essential for protein expression in eukaryotes. RNA-binding proteins (RBPs), central to this regulation, selectively bind to mRNA through RNA-binding domains in a sequence- and structure-specific manner. RBPs orchestrate the processing, storage, and handling of cellular RNAs^7,8 and are involved in cardiovascular biology and disease.⁹ Mitochondrial dysfunction serves as a pivotal pathological mechanism underlying the onset and progression of heart failure.^10,11 Previous studies have demonstrated that GRSF1 (guanine-rich RNA sequence binding factor 1) plays a critical role in maintaining the expression of mitochondrially encoded proteins and preserving mitochondrial homeostasis.¹² Through integrative analysis of heart failure databases and bioinformatic screening for mitochondrial-associated RBPs, we identified GRSF1 as an important regulator of the post-transcriptional events in heart failure. GRSF1, identified as an RBP, belongs to the heterogeneous nuclear ribonucleoprotein F/H subfamily. It is ubiquitously expressed, containing 3 quasi-RNA recognition motifs (qRRMs) that preferentially bind to G-rich sequences within target mRNAs.^13,14 To date, GRSF1 has been implicated in RNA post-transcriptional regulatory events, including RNA transport and localization,¹⁵ RNA stability,^16,17 translation,^18,19 and alternative splicing.^20,21 Apart from the nucleus and cytoplasm, GRSF1 localizes to the mitochondria, in which it regulates diverse biological processes related to mitochondrial metabolism, including mitochondrial ribosome biosynthesis^22,23 and RNA processing and trafficking.^12,15 Functionally, GRSF1 maintains mitochondrial homeostasis, suppresses mitochondrial reactive oxygen species production and pro-inflammatory transcriptional programs, and mitigates age-related hypercoagulability at the cellular level.^17,24,25 Notably, GRSF1 has been captured in long noncoding RNA cytb pull-down assays from the cytosol and mitochondria, in which long noncoding RNA cytb protects pressure overload–induced mice from heart failure.²⁶ This highlights the pivotal role of GRSF1 in maintaining cardiac health. Given the paucity of in vivo studies on GRSF1, the close relationship of GRSF1 and mitochondrial metabolic homeostasis, and the enigma of post-transcriptional regulation of the BCKDH complex, we examined the role of GRSF1 in cardiac BCAA metabolism and its implications for cardiac function.

In this study, we aimed to explore the role of GRSF1 in cardiac BCAA homeostasis using cardiomyocyte-specific GRSF1 or BCKDHB knockout mice. We found that mice with GRSF1 ablation developed heart failure and exhibited severe disruptions of BCAA metabolism. These findings were further corroborated in mice overexpressing cardiomyocyte-specific GRSF1. We also discovered that GRSF1 directly interacted with the G-tracts in the coding region (CR) of BCKDHB mRNA, enhancing BCKDHB mRNA stability at the post-transcriptional level and promoting its protein expression, which was dependent on its qRRMs. Importantly, the cardiac functional improvement mediated by GRSF1 overexpression was partially abolished by BCKDHB knockout. Our findings reveal a novel mechanism for controlling BCAA catabolism in heart failure.

Methods

Human Samples

Human heart samples used in this study were acquired from the Heart Transplant Center of Fuwai Hospital, Beijing, China. The use of human heart samples was approved by the ethics committee of Fuwai Hospital (approval No. 2013-496), with written informed consent obtained from the immediate family of organ donors. The study included 10 patients with heart failure who had undergone heart transplantation surgery because of dilated cardiomyopathy (DCM) as well as 6 healthy samples from brain-dead donors who were not suitable for transplantation for either technical or noncardiac reasons (Table S1).

Animal Studies

All animal studies were conducted following approved protocols by the institutional animal care and use committee at Peking University Health Science Center (approval No. PUIRB-LA2023301), in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Data Availability

The raw data for the cardiac metabolomics were uploaded to the database of MetaboLights (accession MTBLS13178). All supporting data are available from the corresponding author upon reasonable request. A Major Resources Table and detailed methods are provided in the Supplemental Material.

Statistical Analysis

Data were analyzed using the GraphPad Prism software (version 9.4.1) and are presented as the mean±SEM. The Shapiro-Wilk test was used to assess the normality of data distribution. For normally distributed data, group differences were analyzed using a 2-tailed t test for 2-group comparisons or 1- or 2-way ANOVA for comparisons involving multiple groups, followed by the Tukey post hoc test for pairwise comparisons. For non-normally distributed data, the Mann-Whitney U test was used for comparisons between 2 groups and the Kruskal-Wallis test for multiple group comparisons, followed by Dunn’s post hoc analysis for pairwise comparisons. Statistical significance was set at P<0.05.

Results

GRSF1 Expression Decreases in Heart Failure

To identify important RBPs involved in heart failure, we performed an integrated analysis of transcriptomic datasets from both DCM and ischemic cardiomyopathy. Comparative bioinformatics analysis identified 272 differentially expressed RBPs, including well-characterized cardiac regulators such as DDX5²⁷ and QKI,²⁸ which have been previously implicated in heart failure (Figure S1A). Mitochondrial dysfunction serves as a pivotal pathological mechanism underlying the onset and progression of heart failure. To identify novel RBPs associated with mitochondrial metabolism in heart failure, we further performed Venn analyses to screen 57 potential candidate genes (Figure S1B). Gene Ontology enrichment analysis identified 16 RBPs involved in post-transcriptional events, with GRSF1 specifically responsible for mRNA binding (Figure S1C and S1D). Analysis of RNA sequencing data (GSE116250) revealed a significant decrease in GRSF1 mRNA levels (Figure S1E) in left ventricle samples from patients with DCM and ischemic cardiomyopathy. Correlation analysis showed a negative association between GRSF1 expression and the levels of atrial natriuretic peptide (ANP) or brain natriuretic peptide (BNP) in human myocardium (Figure S1F), suggesting a positive role of GRSF1 in regulating cardiac function. GRSF1 is expressed across a broad spectrum, with particularly high levels observed in the heart (Figure S1G). Consistent with the transcriptional data, patients with DCM exhibited reduced GRSF1 protein levels in the myocardium compared with those with nonfailing myocardium (Figure 1A and 1B; Table S1). Correlation analysis further revealed a strong negative correlation between GRSF1 expression and the mRNA levels of ANP, BNP, and β-MHC (β-myosin heavy chain) (Figure 1C and 1D). Additionally, in murine heart failure models induced by transverse aortic constriction (TAC) or isoproterenol treatment, GRSF1 protein levels significantly decreased (Figure 1E through 1H). Notably, GRSF1 protein levels decreased in cardiomyocytes but not in cardiac fibroblasts isolated from adult mice after TAC (Figure 1I). In vitro experiments further showed a significant reduction in GRSF1 protein levels in primary neonatal mouse cardiomyocytes after 24 h of isoproterenol treatment (Figure 1J). These findings suggest that GRSF1 may be implicated in maintaining myocardial homeostasis.

Figure 1.

GRSF1 (guanine-rich RNA sequence binding factor 1) is downregulated in heart failure. A, Representative immunohistochemical images (top, brown) and comparison (bottom) of GRSF1 expression in paraffin-embedded sections from nonfailing human hearts (n=6) and dilated cardiomyopathy (n=10). Scale bars=250 μm and 50 μm. B, Representative Western blot images (left) and comparisons (right) of GRSF1 protein levels in nonfailing human hearts (n=6) and dilated cardiomyopathy (n=10), with GAPDH as a loading control. C, Real-time quantitative analysis showing atrial natriuretic peptide, brain natriuretic peptide, and β-myosin heavy chain mRNA expression in nonfailing human hearts (n=6) and dilated cardiomyopathy (n=10). D, Correlation analysis of GRSF1 protein levels with the mRNA levels of atrial natriuretic peptide, brain natriuretic peptide, and β-myosin heavy chain in human hearts (n=16). E, Representative Western blot images (left) and comparisons (right) of GRSF1 protein levels in murine hearts subjected to 8 weeks of transverse aortic constriction (TAC) to induce heart failure. GAPDH served as a loading control (n=6 per group). F, Representative immunohistochemical images depicting GRSF1 protein levels in murine hearts subjected to 8 weeks of transverse aortic constriction to induce heart failure (n=6 per group). G, Representative Western blot images (left) and comparisons (right) of GRSF1 protein levels in murine hearts subjected to 12 weeks of isoproterenol to induce heart failure, with GAPDH as a loading control (n=6 per group). H, Representative immunohistochemical images depicting GRSF1 protein levels in murine hearts subjected to 12 weeks of isoproterenol to induce heart failure (n=6 per group). I, Representative Western blot images (left) and comparisons (right) of GRSF1 protein levels in adult cardiomyocytes and fibroblasts from mice with transverse aortic constriction–induced heart failure (n=5 per group). J, Representative Western blot images (left) and comparisons (right) of GRSF1 protein levels in primary neonatal mouse cardiomyocytes treated with isoproterenol for 12, 24, and 48 h (n=5 per group). Data are presented as mean±SEM. Statistical significance was assessed using a 2-tailed unpaired t test (A, B, E, and G), 2-tailed nonparametric Mann-Whitney test (C), 2-tailed nonparametric Spearman correlation (D), 1-way ANOVA followed by Tukey post hoc test (J), or 2-way ANOVA followed by Tukey post hoc test (I). β-MHC indicates β-myosin heavy chain; ANP, atrial natriuretic peptide; BNP, brain natriuretic peptide; CF, cardiac fibroblast; CM, cardiomyocytes; DCM, dilated cardiomyopathy; ISO, isoproterenol; ns, not significant; and TAC, transverse aortic constriction.

Cardiac-Specific GRSF1 Deletion Leads to Heart Damage

To investigate the in vivo role of GRSF1 in regulating myocardial function, cardiac-specific GRSF1 (αMHC-GRSF1^fl/fl; GRSF1^△CM) knockout mice were generated (Figure S2A and S2B). The GRSF1^△CM mice were not embryonic lethal, did not exhibit severe growth retardation, and were born at Mendelian ratios. GRSF1 protein was barely expressed in cardiac tissues specifically deleted by cardiomyocytes, further indicating that GRSF1 is highly expressed in cardiomyocytes (Figure 2A). However, mice with deleted GRSF1 had a significantly shorter life span with a median survival age of ≈8 months (Figure 2B). Continuous echocardiography showed that cardiac systolic function of GRSF1 deletion mice exhibited a stunning decline, which was manifested by decreased left ventricular ejection fraction and fraction shortening as well as increased end-diastolic left ventricular internal diameter, indicating dilation of the left ventricular chamber (Figure 2C through 2F). In addition, mice with GRSF1 ablation displayed increased heart weight-to-tibia length ratio, thinned left ventricular wall, increased gross heart sectional area, enlarged myocyte area, and aggravated myocardial fibrosis compared with control mice (Figure 2G through 2K). It also decreased the levels of ANP, BNP, and β-MHC mRNA (Figure 2L). These results suggest that GRSF1 deficiency in cardiomyocytes results in cardiac dysfunction and heart failure.

Figure 2.

GRSF1 (guanine-rich RNA sequence binding factor 1) deficiency causes cardiac dilation and dysfunction. A, Representative Western blot images (left) and comparisons (right) of GRSF1 protein levels in the hearts of GRSF1^fl/fl and GRSF1^△CM mice, with GAPDH as a loading control (n=6 per group). B, Survival curves for GRSF1^fl/fl and GRSF1^△CM mice (n=18–20). The median survival times for GRSF1^△CM mice are indicated. C, Representative left ventricular M-mode echocardiography images in GRSF1^fl/fl and GRSF1^△CM mice at 24 weeks. D, Quantification of left ventricular ejection fraction at 8, 16, and 24 weeks (n=6 per group). E, Quantification of left ventricular fractional shortening at 8, 16, and 24 weeks (n=6 per group). F, Quantification of left ventricular internal diameter at end-diastole at 8, 16, and 24 weeks (n=6 per group). G, Ratios of heart weight-to-tibia length in GRSF1^fl/fl and GRSF1^△CM mice at 24 weeks (n=6 per group). H, Representative images of hematoxylin-eosin, wheat germ agglutin, and Masson trichrome staining of cardiac tissues in GRSF1^fl/fl and GRSF1^△CM mice at 24 weeks. Scale bars=2.5 mm and 50 μm for hematoxylin-eosin, 10 μm for wheat germ agglutin, and 2.5 mm and 100 μm for Masson trichrome staining. I, Quantification of the gross heart sectional area (n=6 per group). J, Quantification of myocyte area by wheat germ agglutin staining (n=6 per group). K, Quantification of cardiac fibrosis area using Masson trichrome staining (n=6 per group). L, RT-qPCR analysis showing atrial natriuretic peptide, brain natriuretic peptide, and β-myosin heavy chain mRNA expression in GRSF1^fl/fl and GRSF1^△CM mice at 24 weeks (n=6 per group). Data are presented as mean±SEM. Statistical significance was evaluated using a 2-tailed unpaired t test. β-MHC indicates β-myosin heavy chain; ANP, atrial natriuretic peptide; BNP, brain natriuretic peptide; EF, ejection fraction; exp, exposure; FS, fractional shortening; HE, hematoxylin-eosin; LVID;d, left ventricular internal diameter at end-diastole; W, week; and WGA, wheat germ agglutinin.

Cardiac-Specific GRSF1 Overexpression Alleviates TAC-Induced Heart Failure

Given the evident cardiac dysfunction observed in GRSF1 knockout mice, we investigated the potential protective role of GRSF1 overexpression against pressure overload–induced heart failure by administrating adeno-associated virus 9 (AAV9), cTNT (cardiac troponin T), GRSF1 via tail vein infusion (Figure 3A). Two weeks after AAV9-cTNT-GRSF1 administration, the expression of GRSF1 in the heart was significantly increased (Figure 3B). The mice then underwent TAC treatment. Cardiac function was dynamically analyzed after surgery. Echocardiographic analysis revealed that TAC treatment could significantly reduce cardiac systolic function, as evidenced by decreased left ventricular ejection fraction and fraction shortening, along with increased left ventricular internal dimensions. However, these parameters were significantly alleviated in GRSF1-overexpressing mice (Figure 3C through 3F). To further elucidate the cardioprotective effects of GRSF1 in TAC-induced heart failure, the morphological and molecular biology of the hearts in the mice were analyzed. TAC-induced heart failure mice exhibited features of cardiac hypertrophy and fibrosis, including increased lung weight-to-tibia length ratio and heart weight–to–tibia length ratio; increased gross heart sectional area; enlarged myocyte area; elevated levels of ANP, BNP, and β-MHC mRNAs; and enhanced collagen deposition; these changes were rescued in GRSF1-overexpressing hearts (Figure 3G through 3M). Therefore, GRSF1 may protect against hemodynamic overload–induced heart failure.

Figure 3.

GRSF1 (guanine-rich RNA sequence binding factor 1) overexpression attenuates transverse aortic constriction (TAC)–induced heart failure. A, Eight-week-old mice were injected via the tail vein with an adeno-associated virus of serotype 9 (AAV9) to overexpress GRSF1 under transcriptional control of the cardiomyocyte-specific troponin T promoter (1.4×10¹¹ vector genomes/mouse). Two weeks after AAV9 injection, the mice underwent TAC surgery and were monitored for 8 weeks. B, Representative Western blot images (left) and comparisons (right) of GRSF1 protein levels in the hearts of AAV-Ctrl and AAV-GRSF1 mice, with GAPDH as a loading control (n=6 per group). C, Representative left ventricular M-mode echocardiography images of AAV-Ctrl and AAV-GRSF1 mice 8 weeks after TAC. D, Quantification of left ventricular ejection fraction in AAV-Ctrl and AAV-GRSF1 mice 4, 6, and 8 weeks after TAC (n=6 per group). E, Quantification of left ventricular fractional shortening in AAV-Ctrl and AAV-GRSF1 mice 4, 6, and 8 weeks after TAC (n=6 per group). F, Quantification of left ventricular internal diameter at end-diastole in AAV-Ctrl and AAV-GRSF1 mice 4, 6, and 8 weeks after TAC (n=6 per group). G, Ratio of wet lung weight-to-tibia length in AAV-Ctrl and AAV-GRSF1 mice 8 weeks after TAC (n=6 per group). H, Representative whole-heart images (top) and quantitative analysis of heart weight-to-tibia length ratio (bottom) in AAV-Ctrl and AAV-GRSF1 mice 8 weeks after TAC (n=6 per group). I, Representative images of hematoxylin-eosin, wheat germ agglutinin, and Masson trichrome staining of cardiac tissues in AAV-Ctrl and AAV-GRSF1 mice 8 weeks after TAC. Scale bars=2.5 mm and 50 μm for hematoxylin-eosin, 10 μm for wheat germ agglutinin, and 2.5 mm and 100 μm for Masson trichrome staining. J, Quantification of the gross heart sectional area (n=6 per group). K, Quantification of myocyte area by wheat germ agglutinin staining (n=6 per group). L, Quantification of the cardiac fibrosis area using Masson trichrome staining (n=6 per group). M, RT-qPCR analysis showing atrial natriuretic peptide, brain natriuretic peptide, and β-myosin heavy chain mRNA expression in AAV-Ctrl and AAV-GRSF1 mice 8 weeks after TAC (n=6 per group). Data are presented as mean±SEM. Statistical significance was evaluated by a 2-tailed unpaired t test (B) or 2-way ANOVA followed by Tukey post hoc test (D through H and J through M). β-MHC indicates β-myosin heavy chain; ANP, atrial natriuretic peptide; BNP, brain natriuretic peptide; cTNT, cardiac troponin T; EF, ejection fraction; FS, fractional shortening; HE, hematoxylin-eosin; LVID;d, left ventricular internal diameter at end-diastole; W, week; and WGA, wheat germ agglutinin.

GRSF1 Ablation Suppresses BCAA Catabolism and Promotes Protein Synthesis

To further explore the mechanisms by which GRSF1 impacts cardiac function, we performed untargeted metabolomics analysis using heart tissues from 6-month-old GRSF1^fl/fl and GRSF1^△CM mice. A total of 791 metabolites were identified, encompassing diverse classes, including lipids, amino acids, carbohydrates, and bile acids. Principal component analysis revealed significant metabolic differences between GRSF1 deletion mice and control mice (Figure 4A). As shown in the volcano plot, a total of 91 metabolites were significantly altered in GRSF1^△CM mice compared with GRSF1^fl/fl mice, of which 47 were decreased and 44 were increased (Figure 4B). Hierarchical clustering heat maps further indicated elevated levels of valine, isoleucine, and leucine in GRSF1 deletion mice (Figure 4C). Metabolomic profiling (Figure S3A and S3B) was also analyzed by using cardiac tissues from 3-month-old GRSF1^fl/fl and GRSF1^△CM mice (before onset of significant cardiac dysfunction). Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis showed that differentially expressed metabolites were predominantly enriched in the BCAA metabolism pathway, which ranked as the top pathway (Figure S3C). To clarify the role of GRSF1 in BCAA metabolism, we performed targeted metabolomics to detect BCAAs, BCKAs, and citric acid cycle metabolites. Mice with GRSF1 deletion showed accumulation of valine, isoleucine, and leucine in the heart (Figure 4D). Additionally, the levels of α-ketoisovalerate, α-ketoisocaproate, and α-keto-β-methylvalerate, the catabolites of BCAAs, were significantly elevated, whereas citric acid cycle intermediates such as acetyl coenzyme A, succinyl coenzyme A, succinate, malate, and glutamate were significantly decreased (Figure 4E and 4F). To determine whether GRSF1 directly regulates myocardial BCAA metabolism and that this regulation is not solely an outcome of heart failure, the levels of BCAA were dynamically monitored. Notably, whereas comparable BCAA levels were observed at 8 weeks of age, GRSF1^△CM mice exhibited altered BCAA metabolism by 12 weeks (before onset of significant cardiac dysfunction), with these metabolic differences becoming more pronounced at 16 weeks of age (Figure S3D). These results suggest that GRSF1 deficiency may serve as a primary driver for myocardial BCAA accumulation.

Figure 4.

GRSF1 (guanine-rich RNA sequence binding factor 1) deletion inhibits branched-chain amino acid catabolism and aggravates mitochondrial dysfunction. A, Principal component analysis of untargeted metabolomics in the heart of GRSF1^fl/fl and GRSF1^△CM mice at 24 weeks (n=10 per group). B, Volcano chart of detected metabolites in GRSF1^fl/fl and GRSF1^△CM mice. The 3 metabolites specifically indicated in the plots are valine, isoleucine, and leucine. C, Heat map depicting the top 20 differentially abundant metabolites in GRSF1^fl/fl and GRSF1^△CM mice, identified by a variable importance in projection score >1 and P<0.05 and subsequently ranked by ascending P value. Red arrows indicate significantly elevated levels of branched-chain amino acids. The relative magnitude and distribution of the data are depicted by a color scale, on which the transition in hue from blue to red represents an increase in value. D, Targeted metabolomics analysis focusing on branched-chain amino acids s in GRSF1^fl/fl and GRSF1^△CM mice (n=10 per group). E, Targeted metabolomics analysis focusing on branched-chain keto acids in GRSF1^fl/fl and GRSF1^△CM mice (n=10 per group). F, Targeted metabolomics analysis focusing on energy metabolism in GRSF1^fl/fl and GRSF1^△CM mice (n=10 per group). G, Representative Western blot images (top) and comparisons (bottom) of puromycin incorporation into newly synthesized proteins in the hearts of GRSF1^fl/fl and GRSF1^△CM mice at 24 weeks, with GAPDH as a loading control (n=6 per group). H, Representative Western blot images showing phospho (p)-ERK1/2 (extracellular regulated protein kinase 1/2), ERK1/2, p-mTOR (mammalian target of rapamycin), mTOR, p-4E-BP1 (4E-binding protein 1), and 4E-BP1 in the hearts of GRSF1^fl/fl and GRSF1^△CM mice at 24 weeks, with GAPDH as a loading control. I, Quantification of the p-4E-BP1/4E-BP1 ratio (n=6 per group). J, Quantification of the p-ERK1/2/ ERK1/2 ratio (n=6 per group). K, Quantification of the p-mTOR/mTOR ratio (n=6 per group). L, Electron microscopy images showing cardiac mitochondrion morphology in GRSF1^fl/fl and GRSF1^△CM mice. M, Quantification of the average mitochondrion size, mitochondrion numbers, and crista numbers from L (n=6 per group). N, Assessment of mitochondrial oxidative respiration (left) and quantitative analysis of basal respiration and maximal respiration (right) in adult cardiomyocytes from GRSF1^fl/fl and GRSF1^△CM mice by measuring cellular oxygen consumption rate (n=6 per group). Data are presented as mean±SEM. Statistical significance was evaluated using a 2-tailed unpaired t test (G, I through K, and N) or 2-tailed nonparametric Mann-Whitney test (D through F and M). α-KIC indicates α-ketoisocaproate; α-KIV, α-ketoisovalerate; α-KMV, α-keto-β-methylvalerate; BCAA, branched-chain amino acid; BCKA, branched-chain keto acid; CoA, coenzyme A; and OCR, oxygen consumption rate.

Given that the reamination of BCAAs and BCKAs promotes protein synthesis and plays a key role in cardiac remodeling,^29,30 we hypothesized that protein translation could be altered in GRSF1 deletion mice. The incorporation of puromycin into newly synthesized proteins was significantly higher in heart tissue from 6-month-old GRSF1^△CM mice compared with GRSF1^fl/fl mice (Figure 4G). It has been reported that BCAA-mediated total protein translation is driven by phosphorylation of the eukaryotic translation initiation factor 4E-BP1 (4E-binding protein 1).^31,32 GRSF1 deletion significantly increased in 4E-BP1 phosphorylation, which activated protein synthesis by deactivating the translation repressor function (Figure 4H and 4I). Furthermore, phosphorylation of ERK1/2 (extracellular regulated protein kinase 1/2) enhances the activity of mTOR (mammalian target of rapamycin), which promotes 4E-BP1 phosphorylation.^33,34 GRSF1^△CM mice exhibited elevated phosphorylation levels of ERK1/2 and mTOR (Figure 4H, 4J, and 4K). To further characterize the temporal regulation of protein synthesis by GRSF1, we conducted comparative puromycin incorporation assays in GRSF1^fl/fl and GRSF1^△CM mice. This revealed an age-dependent progression in protein synthesis enhancement. Whereas equivalent levels of nascent proteins were observed at 8 weeks of age, GRSF1^△CM mice displayed progressively increased protein synthesis at 12 weeks of age, with further augmentation by 16 weeks of age (Figure S3E). These results were also consistently reflected in the activation patterns of associated signaling pathways (Figure S3F through S3I).

Considering that mitochondria are key sites of BCAA catabolism, we analyzed the mitochondrial ultrastructure and function. Transmission electron microscopy showed abnormal mitochondrial morphology in GRSF1^△CM mice, characterized by reduced mitochondrion size, increased mitochondrion numbers, and disorganized mitochondrion cristae (Figure 4L and 4M). Seahorse analysis of mitochondrial respiratory capacity revealed a dramatic reduction in both basal and maximal respiratory capacity, indicating impaired mitochondrial function after GRSF1 deletion (Figure 4N). Additionally, mitochondrial reactive oxygen species levels were elevated, and membrane potential levels were significantly reduced in GRSF1-deficient cardiomyocytes (Figure S4A through S4D). Collectively, these results suggest that GRSF1 inhibits BCAA-mediated protein synthesis and mitochondrial dysfunction in cardiomyocytes.

GRSF1 Regulates the Turnover of BCKDHB mRNA

To investigate the mechanisms underlying GRSF1-associated BCAA catabolic disorders in heart failure, we evaluated the role of GRSF1 in regulating the expression of the BCKDH complex, a rate-limiting step that controls BCAA catabolic metabolism (Figure 5A). The BCKDH complex comprises 3 subunits: E1 (encoded by the BCKDHA and BCKDHB genes), E2 (encoded by the DBT gene), and E3 (encoded by the DLD gene). GRSF1 knockdown in cardiomyocytes significantly reduced the protein levels of BCKDHB, whereas the protein expression of BCKDHA, DBT, and DLD remained unchanged (Figure 5B). Accordingly, the mRNA levels of BCKDHB were reduced in neonatal mouse cardiomyocytes when GRSF1 was silenced (Figure 5C). However, short knockdown of GRSF1 did not affect the BCKDHB pre-mRNA levels or translation efficiency in cells cultured in vitro (Figure 5D; Figure S5A through S5D). These results suggest that GRSF1 may regulate the turnover of BCKDHB mRNA. By using RNA pull-down assays, GRSF1 was found to associate with BCKDHB mRNA at the CR but not at the 5′ untranslated region or 3′ untranslated region (Figure 5E and 5F). We further split the full-length CR of BCKDHB mRNA into 6 fragments and performed RNA pull-down assays, in which fragment 4 of the BCKDHB mRNA CR was the primary binding region for GRSF1. Fragment 4 contains a G-rich sequence motif, which likely facilitates binding to GRSF1. To investigate the contribution of the G-rich sequence in association with GRSF1, we either deleted the G-rich sequence or substituted the middle G with U in each G-tract. The RNA pull-down assays showed that G-rich sequence deletion or mutation almost completely abolished the ability to bind to GRSF1 (Figure 5E and 5G). The association of GRSF1 with BCKDHB mRNA was further confirmed by ribonucleoprotein immunoprecipitation assays (Figure 5H). In primary neonatal mouse cardiomyocytes silenced with GRSF1, the half-life of BCKDHB mRNA was greatly shortened (Figure 5I). In addition, knockdown of GRSF1 significantly reduced the activity of the pGL3-derived reporter bearing the CR of BCKDHB mRNA but not that bearing the 5′ untranslated region or 3′ untranslated region of BCKDHB mRNA or those with a deleted or mutated CR of BCKDHB mRNA (Figure 5J). Importantly, the protein and mRNA levels of BCKDHB were also reduced in the heart tissues from GRSF1^△CM mice (Figure 5K and 5L). Furthermore, a reduction of BCKDHB was observed in the hearts of humans with DCM and mouse models of pressure overload–induced heart failure (Figure 5M; Figure S6A and S6B). Correlation analysis revealed that the GRSF1 expression correlated positively with BCKDHB in human myocardium (Figure 5N). Moreover, the level of BCAAs increased significantly in the hearts of patients with DCM and is negatively correlated with GRSF1 protein expression (Figure S6C and S6D). Therefore, GRSF1 promotes the expression of BCKDHB.

Figure 5.

GRSF1 (guanine-rich RNA sequence binding factor 1) directly interacts with the BCKDHB (branched-chain keto acid dehydrogenase E1 subunit β) mRNA coding region (CR) to preserve its stability and protein expression. A, A schematic of the major steps and enzyme complexes involved in the branched-chain amino acid catabolic pathway. B, Representative Western blot images (left) showing GRSF1, BCKDHA, BCKDHB, dihydrolipoyl transacylase, and dihydrolipoamide dehydrogenase and comparisons (right) of BCKDHB protein levels in primary neonatal mouse cardiomyocytes (NMCMs) treated with adenovirus-silencing GRSF1, with GAPDH as a loading control (n=6 per group). C, RT-qPCR analysis of BCKDHB mRNA expression in primary NMCMs treated with adenovirus-silenced GRSF1 (n=6 per group). D, RT-qPCR analysis of BCKDHB pre-mRNA expression in primary NMCMs treated with adenovirus-silenced GRSF1 (n=6 per group). E, Schematic diagram of BCKDHB mRNA 5′ untranslated region (UTR), CR, and 3′ UTR; different fragments of the BCKDHB mRNA CR; and position of the G-rich sequences in the BCKDHB mRNA CR showing the wild-type plasmid, deleted-type plasmid, and mutated plasmid. F, Biotin-labeled fragments of BCKDHB mRNA 5′ UTR, CR, and 3′ UTR were amplified and subjected to RNA pull-down assays to detect GRSF1 by Western blot. G, Biotin-labeled fragments of different regions of the BCKDHB mRNA CR (left) and the BCKDHB mRNA deleted or mutated CR (right) were amplified and subjected to RNA pull-down assays to detect GRSF1 by Western blot. H, Ribonucleoprotein immunoprecipitation assays were performed using Immunoglobulin G or anti-GRSF1 antibodies in primary NMCMs. RT-qPCR analysis shows BCKDHB and GAPDH mRNA expression (n=6 per group). I, Primary NMCMs infected with an adenovirus to silence GRSF1 for 48 hours were evaluated for the half-lives of BCKDHB and GAPDH mRNA using actinomycin D (n=6 per group). J, Primary NMCMs were infected with an adenovirus to silence GRSF1 for 24 hours and further transfected with each of the pGL3-derived reporters along with a renilla luciferase–thymidine kinasevector and cultured for an additional 48 hours. Relative luciferase activities were then determined (n=6 per group). K, Representative Western blot images (left) showing BCKDHA, BCKDHB, dihydrolipoyl transacylase, and dihydrolipoamide dehydrogenase and comparisons (right) of BCKDHB protein levels in the hearts of GRSF1^fl/fl and GRSF1^△CM mice, with GAPDH as a loading control (n=6 per group). L, RT-qPCR analysis showing BCKDHB mRNA expression in the hearts of GRSF1^fl/fl and GRSF1^△CM mice (n=6 per group). M, Representative Western blot images (left) and comparisons (right) of BCKDHB protein levels in nonfailing human hearts (n=6) and dilated cardiomyopathy (n=10). N, Correlation analysis between GRSF1 and BCKDHB protein levels in the human heart (n=16). Data are presented as mean±SEM. Statistical significance was evaluated using a 2-tailed unpaired t test (B through D, H, and J through M) or 2-tailed nonparametric Spearman correlation (N). α-KG indicates α-Ketoglutarate; α-KIC, α-ketoisocaproate; α-KIV, α-ketoisovalerate; α-KMV, α-keto-β-methylvalerate; ActD, actinomycin D; BCAA, branched-chain amino acid; BCKA, branched-chain keto acid; BCKDHA, branched chain keto acid dehydrogenase E1 subunit α; BCKDHB, branched chain keto acid dehydrogenase E1 subunit β; CoA, coenzyme A; CR, coding region; DBT, dihydrolipoyl transacylase; DCM, dilated cardiomyopathy; DEL, deleted; DLD, dihydrolipoamide dehydrogenase; Glu, glutamic acid; ns, not significant; IgG, Immunoglobulin G; MUT, mutated; sh-NC, negative control; TCA, tricarboxylic acid cycle; UTR, untranslated region; and WT, wild type.

The qRRM Domains Are Required for GRSF1 to Regulate BCKDHB

GRSF1 contains qRRM domains that primarily bind to the G-rich RNA sequences. To determine the specific qRRM domain required to interact with the BCKDHB mRNA CR, we generated truncated mutants lacking the qRRM domain and performed RNA pull-down assays. Compared with the wild-type (WT) GRSF1, the ∆qRRM2 truncation completely lost the ability to bind to the BCKDHB mRNA CR, whereas the ∆qRRM1 and ∆qRRM3 truncations retained partial binding activity, implying that the qRRM2 domain is bona fide required for GRSF1 binding to the BCKDHB mRNA CR (Figure 6A). Taken together, GRSF1 promotes BCKDHB expression by stabilizing BCKDHB mRNA by binding GRSF1 qRRM2 to the G-rich sequence of the BCKDHB mRNA CR.

Figure 6.

The functional role of GRSF1 (guanine-rich RNA sequence binding factor 1) depends critically on its quasi-RNA recognition motif 2 (qRRM2) domain. A, Schematic of the 3 qRRM domains of GRSF1 in mice, along with their truncations lacking qRRM1, qRRM2, or qRRM3 domains. Biotin-labeled fragments of the BCKDHB (branched-chain keto acid dehydrogenase E1 subunit β) mRNA coding region were amplified and subjected to an RNA pull-down assay to detect Flag-GRSF1 by Western blot. B, Six-week-old GRSF1^△CM mice were injected via a tail vein with adeno-associated virus serotype 9 (AAV9)–cardiac troponin T–GRSF1 wild-type (WT) or AAV9-cTNT-GRSF1 mutated to generate GRSF1 WT or GRSF1-∆qRRM2 truncation mice. Two weeks after AAV9 injection, the mice underwent transverse aortic constriction surgery and were monitored for 6 weeks. Representative Western blot images (top) and comparisons (bottom) of GRSF1 and BCKDHB expression in murine hearts infected with AAV-GRSF1 WT or AAV-GRSF1 mutated, with GAPDH as a loading control (n=6 per group). C, Targeted metabolomics analysis of branched-chain amino acids in GRSF1^△CM mice treated with AAV-GRSF1 WT or AAV-GRSF1 mutated 6 weeks after transverse aortic constriction (n=8 per group). D, Targeted metabolomics analysis of branched-chain keto acids in GRSF1^△CM mice treated with AAV-GRSF1 WT and AAV-GRSF1 mutated 6 weeks after transverse aortic constriction (n=8 per group). E, Representative left ventricular M-mode echocardiography images 6 weeks after transverse aortic constriction. F, Quantification of left ventricular ejection fraction (n=6 per group). G, Quantification of left ventricular fractional shortening (n=6). H, Quantification of left ventricular internal diameter at end-diastole and left ventricular internal diameter at end-systole (n=6 per group). I, Ratio of wet lung weight-to-tibia length (n=6 per group). J, Representative whole-heart images (left) and quantitative analysis of heart weight-to-tibia length ratios (right, n=6 per group). K, Representative images of hematoxylin-eosin, wheat germ agglutinin, and Masson trichrome staining of cardiac tissues. Scale bars=2.5 mm and 50 μm for hematoxylin-eosin, 10 μm for wheat germ agglutinin, and 2.5 mm and 100 μm for Masson trichrome staining. L, Quantification of the gross heart sectional area (n=6 per group). M, Quantification of myocyte area by wheat germ agglutinin staining (n=6 per group). N, Quantification of cardiac fibrosis area by Masson trichrome staining (n=6 per group). Data are presented as mean±SEM. Statistical significance was evaluated by 1-way ANOVA followed by Tukey post hoc test (B through D, I, J, and L through N) or Kruskal-Wallis test followed by Dunn post hoc test (F through H). α-KIC indicates α-ketoisocaproate; α-KIV, α-ketoisovalerate; α-KMV, α-keto-β-methylvalerate; BCAA, branched-chain amino acid; BCKA, branched-chain keto acid; EF, ejection fraction; FS, fractional shortening; HE, hematoxylin-eosin; LVID;d, left ventricular internal diameter at end-diastole; LVID;s, left ventricular internal diameter at end-systole; MUT, mutated; and WGA, wheat germ agglutinin.

To investigate the role of GRSF1 qRRM2 in vivo, we generated GRSF1 WT and GRSF1-∆qRRM2 truncation mice by administering AAV9-cTNT-GRSF1 WT or AAV9-cTNT-GRSF1 MUT (∆qRRM2 truncation) via tail vein infusion in GRSF1^△CM mice. Western blot analysis revealed that cardiac-specific GRSF1 expression significantly increased BCKDHB protein levels, whereas there was no significant increase in BCKDHB expression in the hearts of GRSF1-ΔqRRM2 truncation mice (Figure 6B). GRSF1 WT reduced the levels of valine, leucine, and isoleucine as well as the levels of α-ketoisocaproate, α-ketoisovalerate, and α-keto-β-methylvalerate compared with AAV-Ctrl (empty vector control), whereas this beneficial effect was only partially observed in GRSF1 MUT mice (Figure 6C and 6D). Additionally, GRSF1 WT mice showed improved cardiac function, reduced cardiac hypertrophy, and attenuated cardiac fibrosis. In contrast, GRSF1 MUT mice showed persistently reduced left ventricular ejection fraction and fractional shortening, along with increased end-diastolic left ventricular internal diameter (Figure 6E through 6H). Additionally, these mice exhibited sustained increases in the lung weight-to-tibia length ratio and heart weight-to-tibia length ratio, left ventricular wall thinning, enlarged gross heart sectional area, cardiomyocyte hypertrophy, and exacerbated myocardial fibrosis (Figure 6I through 6N). These results suggest that the cardioprotective effects of GRSF1 are dependent on the qRRM2 domain.

GRSF1 Exerts Cardioprotective Effects by Regulating BCKDHB

To delineate the temporal progression of cardiac function, we conducted a longitudinal study comparing baseline cardiac function between BCKDHB^fl/fl and BCKDHB^△CM mice (Figure S7A). At 8 weeks of age, BCKDHB^△CM mice displayed cardiac function comparable to that of BCKDHB^fl/fl mice. However, prolonged observation revealed a gradual decline in cardiac function by 14 weeks (preceding the onset of GRSF1^△CM phenotypes), with clear signs of heart failure evident by 20 weeks (Figure S7C through S7G). To investigate whether cardiac-specific GRSF1 expression improves TAC-induced heart failure through increased BCKDHB expression, 6-week-old cardiac-specific BCKDHB knockout mice were further infected with either an AAV expressing GRSF1 or a control virus for 2 weeks. After confirming the effect of the infection, we analyzed the effect of GRSF1 expression in improving TAC-induced heart failure. As anticipated, ablation of BCKDHB significantly reduced cardiac systolic function, indicated by decreased left ventricular ejection fraction and fraction shortening, and increased left ventricular internal diameter (Figure 7A through 7D). Although the expression of GRSF1 significantly improved TAC-induced heart failure in BCKDHB^fl/fl mice, it only produced a weaker effect of recovering the deteriorated cardiac function observed in TAC-induced BCKDHB^△CM mice (Figure 7A through 7D). The changes of the morphology and molecular biology of the hearts from these mice were further analyzed. GRSF1 overexpression mice exhibited features of improved cardiac hypertrophy and fibrosis, including reduced lung and heart weight-to-tibia length ratio; decreased gross heart sectional area and myocyte area; downregulated levels of ANP, BNP, and β-MHC mRNAs; and lower collagen deposition; these changes were significantly attenuated in BCKDHB deletion hearts (Figure 7E through 7K). In addition, we infected the GRSF1^fl/fl and GRSF1^△CM mice with a lesser amount of the viruses to restore BCKDHB expression (Figure S8A) and evaluated the effect of BCKDHB expression on rescuing the phenotypes resulting from GRSF1 ablation. BCKDHB expression effectively mitigated TAC-induced heart failure and restored cardiac function in GRSF1^△CM mice (Figure S8B through S8D). Histological and molecular analysis demonstrated that BCKDHB expression alleviated cardiac hypertrophy and fibrosis phenotypes (Figure S8E through S8I) in TAC-treated GRSF1^△CM mice. These findings suggest that the cardioprotective effects of GRSF1 depend on the expression of BCKDHB.

Figure 7.

BCKDHB (branched-chain keto acid dehydrogenase E1 subunit β) deletion attenuates the cardioprotective effects of GRSF1 (guanine-rich RNA sequence binding factor 1) overexpression in mice with transverse aortic constriction (TAC)–induced heart failure. A, Six-week-old BCKDHB^fl/fl and BCKDHB^△CM mice were injected via a tail vein with adeno-associated virus serotype 9to overexpress GRSF1 under transcriptional control of the cardiomyocyte-specific troponin T promoter (1.4×10¹¹ vector genomes/mouse). Two weeks after adeno-associated virus serotype 9 injection, the mice underwent TAC surgery and were monitored for 6 weeks. Representative left ventricular M-mode echocardiography images 6 weeks after TAC are shown. B, Quantification of left ventricular ejection fraction (n=6 per group). C, Quantification of left ventricular fractional shortening (n=6 per group). D, Quantification of left ventricular internal diameter at end-diastole and LVID;s (n=6 per group). E, Ratio of wet lung weight-to-tibia length 6 weeks after TAC (n=6 per group). F, Representative whole heart images (top) and quantitative analysis of heart weight-to-tibia length ratios (bottom) 6 weeks after TAC (n=6 per group). G, Representative images of hematoxylin-eosin, wheat germ agglutinin, and Masson trichrome staining of cardiac tissues. Scale bars=2.5 mm and 50 μm for hematoxylin-eosin, 10 μm for wheat germ agglutinin, and 2.5 mm and 100 μm for Masson trichrome staining. H, Quantification of the gross heart sectional area (n=6 per group). I, Quantification of myocyte area by wheat germ agglutinin staining (n=6 per group). J, Quantification of cardiac fibrosis area by Masson trichrome staining (n=6 per group). K, RT-qPCR analysis of atrial natriuretic peptide, brain natriuretic peptide, and β-myosin heavy chain mRNA expression levels (n=6 per group). Data are presented as mean±SEM. Statistical significance was evaluated using 2-way ANOVA followed by Tukey post hoc test. β-MHC indicates β-myosin heavy chain; ANP, atrial natriuretic peptide; BNP, brain natriuretic peptide; EF, ejection fraction; FS, fractional shortening; HE, hematoxylin-eosin; LVID;d, left ventricular internal diameter at end-diastole; LVID;s, left ventricular internal diameter at end-systole; and WGA, wheat germ agglutinin.

Discussion

Perturbations of cardiac substrate metabolism, particularly those involving a BCAA metabolic imbalance, are major contributors to cardiac remodeling and heart failure.³⁵ Restoring BCAA metabolic homeostasis is a promising therapeutic strategy for heart failure. This study delves into the role of GRSF1, a known RBP, in regulating cardiac BCAA metabolism and its impact on heart failure progression. Our findings indicate that GRSF1 expression is decreased in the hearts of patients with DCM and failing murine hearts. GRSF1 acts as a crucial endogenous regulator of cardiac substrate metabolism at the post-transcriptional level, which is essential for maintaining BCAA homeostasis and mitochondrial function. Cardiomyocyte-specific GRSF1-deficient mice exhibited reduced cardiac function, progression to spontaneous DCM and heart failure, and increased cardiac remodeling. Further investigation revealed that GRSF1 promoted mRNA stability by directly binding to the G-rich element in the CR of BCKDHB mRNA, a major subunit of the BCKDH complex, thereby preserving BCAA homeostasis. Conversely, increased GRSF1 expression protected cardiomyocytes from TAC-induced BCAA metabolic imbalance and heart failure. Our findings underscore the role of GRSF1 in mediating BCAA homeostasis via BCKDHB, which offers strong protection against the development of heart failure (Figure S9).

Cellular processes mediated by RBPs through RNA splicing, degradation, stabilization, modification, and translation are crucial in the development of cardiovascular diseases, including atherosclerosis, ischemic heart disease, and heart failure.⁹ GRSF1, an RBP localized in the cytoplasm and mitochondria, binds to G-rich regions to regulate mRNA stability and translation.¹³ GRSF1 accumulates at discrete foci in the mitochondrial matrix and plays a role in mitochondrial RNA processing.^12,22 Post-transcriptional regulation of cardiometabolic enzymes in the mitochondria has a significant impact on cardiac metabolism and function. However, the role of GRSF1 in the regulation of cardiac substrate metabolism and its effects on cardiac function are not well understood. To address this, we generated mice with cardiomyocyte-specific GRSF1 deletion. These mice exhibited progressive contractile dysfunction, left ventricular dilatation, increased cardiomyocyte size, severe cardiac fibrosis, and a significantly shortened life span in the absence of GRSF1.

Disturbances of substrate metabolism are intimately involved in the pathogenesis of cardiovascular disease; therefore, metabolic interventions for heart failure have been drawing increasing attention.³⁶ BCAAs, including leucine, isoleucine, and valine, are essential amino acids primarily derived from the diet, though they can also be produced de novo by the gut microbiota.⁵ In both human and animal models of heart failure, the BCAA metabolic pathway is consistently downregulated in the cardiac tissue, leading to accumulation of BCAAs and BCKAs, which contribute to heart failure.^5,37,38 Our study reveals a previously unappreciated role of GRSF1 in mediating cardiac BCAA metabolism. GRSF1 deficiency strongly contributed to the accumulation of cardiac BCAAs and BCKAs. Previous research has demonstrated that impaired cardiac BCAA metabolism is associated with increased cardiac protein synthesis and mitochondrial dysfunction, independent of BCAA use as an energy substrate for the heart.^29,39 Our results indicate that the phosphorylation levels of ERK1/2, mTOR, and 4E-BP1 were significantly elevated in GRSF1 deletion mice, suggesting elevated protein synthesis. Additionally, we observed significant dysregulation of citric acid cycle metabolites in the hearts of GRSF1-deficient mice, potentially because of abnormalities in protein synthesis and mitochondrial function. These findings suggest that GRSF1 plays a crucial role in maintaining both BCAA and mitochondrial homeostasis in cardiomyocytes. However, given that GRSF1 can impact mitochondrial function by regulating mitochondrial RNA processing and is involved in regulation of the ferroptosis-related protein glutathione peroxidase 4,⁴⁰ GRSF1-mediated post-transcriptional regulatory events may contribute to heart failure through mechanisms beyond BCAA metabolism. In addition, whether GRSF1- or BCKDHB-mediated regulation of mitochondrial dysfunction could exert a feedback effect in BCAA metabolism remains to be further studied.

BCAAs are obtained exclusively from dietary sources, necessitating tight regulation during the catabolic phase. The initial step in BCAA catabolism involves the reversible conversion of BCAA to BCKA via branched-chain aminotransferases. Subsequently, BCKAs undergo irreversible oxidization by the BCKDH complex, the rate-limiting step in the degradation pathway, eventually degrading to acetyl coenzyme A and succinyl coenzyme A.⁴¹ The expression and activity of the BCKDH complex are tightly regulated to maintain BCAA homeostasis. The BCKDH complex comprises E1, an α-ketoacid decarboxylase containing the subunits BCKDHA and BCKDHB; E2, a dihydrolipoyl transacylase; and E3, a dihydrolipoamide dehydrogenase.^41,42 Previous studies have shown that mutations in the BCKDHB gene result in accumulation of BCAAs and their corresponding BCKAs, ultimately manifesting as maple syrup urine disease, a metabolic disorder clinically characterized by recurrent, life-threatening neurological crises and progressive cerebral damage.^43–45 Notably, aligning with accumulating evidence,^4,46 BCKDHB expression was significantly downregulated in heart failure, indicating a compelling link between BCAA metabolism and cardiac health. Our findings further indicated that GRSF1 regulates BCKDHB expression in heart failure by stabilizing BCKDHB mRNA, a process dependent on both the structural integrity of GRSF1 and the presence of a G-rich element in the BCKDHB CR. In addition, elevated protein translation in GFRS1 knockout mice >8 weeks of age suggests involvement of translational regulation of BCKDHB. However, this translational upregulation is absent in both in vitro–cultured cells with short GRSF1 knockdown and in young GRSF1 knockout mice <8 weeks of age, indicating that the enhanced protein translation observed in GRSF1-deficient mice likely represents a delayed compensatory mechanism. This process appears to depend on progressive accumulation of BCAAs and BCKAs to ultimately trigger activation of the ERK1/2–mTOR–4E-BP1 signaling axis.^31,34 Critically, concurrent reductions in GRSF1 and BCKDHB proteins were observed in human heart failure, TAC-induced murine models, and GRSF1 knockout mice, demonstrating that ERK1/2–mTOR–4E-BP1 activation cannot compensate for BCKDHB protein loss driven by mRNA destabilization upon GRSF1 depletion. It is important to recognize that our study is constrained by a limitation inherent to most RBP research: the pervasive multitarget nature of GRSF1. The phenotypic consequences observed after GRSF1 manipulation likely represent the integrated outcome of its dysregulation across numerous mRNA targets. Consequently, whereas our data point to a functional interaction between GRSF1 and BCKDHB mRNA, the extent to which BCKDHB specifically mediates the observed phenotypic effects of GRSF1 requires further investigation. It is probable that additional GRSF1 targets contribute to the overall cardiac phenotype.

The present study underscores the significance of the GRSF1-BCKDHB regulatory process for protecting against heart failure. It can be also postulated that targeting dysfunctional BCAA metabolic homeostasis in heart failure may be a promising strategy to improve functional impairments in cardiometabolic disease. Beyond heart failure, dysregulation of BCAA metabolism has been implicated in a broad spectrum of disorders, including atherosclerosis, hypertension, diabetes, and inborn errors of metabolism, such as maple syrup urine disease.^5,6,44 In addition, it should be noted that the exclusive use of male mice in this study constitutes a limitation. Therefore, further investigations involving both sexes are necessary to elucidate whether the GRSF1-BCKDHB regulatory axis influences BCAA homeostasis in these disease contexts.

Article Information

Author Contributions

Drs H. Wang, J. Wang, L. Jin, J. Yang, M. Cui, J. Song, W. Wang, and M. Xu designed the study and wrote the manuscript. Drs H. Wang, M. Zhu, H. Cui, C. Liu, C. Fan, and H. Li performed the experiments and analyzed the data. Dr Song provided the human samples. Drs Xu and W. Wang supervised the study and generated project resources. All authors reviewed and approved the final version of the manuscript.

Sources of Funding

This work was supported by the National Key Research and Development Program of China (2020YFA0803800 and 2020YFA0803803 to Dr M. Xu and 2024YFA0918700 to Dr W. Wang), the Beijing Natural Science Foundation (L248019 to Dr M. Xu), the Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences (2021-I2M-5-003 to Dr M. Xu), and the Research Project of Peking University Third Hospital in State Key Laboratory of Vascular Homeostasis and Remodeling (Peking University, 2025-VHR-SY-09 to Dr M. Xu). Dr H. Wang was supported in part by a postdoctoral fellowship from the Peking-Tsinghua Center for Life Sciences.

Disclosures

None.

Supplemental Material

Checklist

Expanded Methods

Figures S1–S9

Table S1