Circular RNAs (circRNAs) are non-coding RNA molecules with a circular structure that distinguishes them from conventional linear mRNAs. Unlike mRNAs, which have 5' and 3' ends, circRNAs are formed through back-splicing, in which base pairing within inverted repeat elements, such as Alu sequences, brings a downstream splice donor site into proximity with an upstream splice acceptor site to form a covalently closed circular molecule^[1]. Their complex biogenesis may also be influenced by specific RNA-binding proteins (RBPs), including fusion sarcoma protein (FUS), nuclear factor 90 (NF90), and NF110, which bind to intronic regions and promote circRNA formation^[2]. For many years, circRNAs were considered “useless” transcripts arising from splicing errors. However, circRNA-specific identification algorithms and high-throughput RNA-seq have revealed their important roles in gene regulation and other cellular functions. circRNAs lack 5′ caps and 3′ poly(A) tails and are more resistant to RNA exonucleases, more stable, more conserved, and longer-lived than linear RNAs^[3,4]. They may therefore serve as biomarkers and therapeutic targets in disease, particularly in heart failure initiation and progression, although further validation is required. Based on left ventricular ejection fraction, heart failure (HF) is classified as heart failure with reduced ejection fraction (HFrEF), heart failure with mildly reduced ejection fraction (HFmrEF), or heart failure with preserved ejection fraction (HFpEF). These three subtypes differ in clinical phenotype, ventricular structure, major pathophysiological mechanisms, and treatment response. HFrEF primarily involves ventricular systolic dysfunction, whereas HFpEF is typically associated with diastolic dysfunction, systemic inflammation, endothelial and microvascular dysfunction, metabolic comorbidities, and increased myocardial stiffness. However, the roles of circRNAs in HFpEF and HFmrEF remain incompletely understood, despite the partially overlapping pathophysiological features of these HF phenotypes.

Most experimental evidence on circRNAs in heart failure currently comes from models of myocardial infarction, pressure overload, myocardial hypertrophy, and ventricular remodeling. This review therefore focuses on circRNA-mediated molecular mechanisms associated with ventricular dysfunction and remodeling, particularly those relevant to HFrEF-like pathological processes. In contrast, evidence on the roles of circRNAs in HFpEF and HFmrEF remains limited, warranting further investigation.

According to the "General Definition and Classification of Heart Failure" and current guidelines from the American Heart Association/American College of Cardiology/American Heart Failure Society (AHA/ACC/HFSA) and the European Society of Cardiology, HF is a clinical syndrome characterized by symptoms and/or signs caused by structural or functional cardiac abnormalities and supported by objective evidence, such as elevated natriuretic peptide levels or pulmonary and systemic edema.

Heart failure can manifest as acute heart failure (AHF) or chronic heart failure (CHF). AHF is a clinical state in which symptoms and signs develop rapidly or worsen substantially, requiring urgent assessment and treatment. Although AHF and CHF differ considerably in clinical presentation and disease progression, both involve complex molecular pathological mechanisms, including inflammation, cardiomyocyte injury, fibrosis, myocardial hypertrophy, and ventricular remodeling. HF represents a major public health burden worldwide, affecting more than 64 million individuals, with prevalence increasing markedly with age. In Europe and North America, its estimated prevalence in adults is approximately 1%-2%, rising to more than 10% among individuals older than 70 years. Population-based studies in China have also shown an increasing prevalence of HF, largely driven by population aging and the growing burden of hypertension, coronary artery disease, diabetes, and obesity. Despite therapeutic advances, HF remains associated with high mortality and frequent hospitalization, particularly in patients with advanced disease or acute decompensation. These epidemiological data highlight the need for improved molecular biomarkers and therapeutic targets^[5-8]. In AHF, severe myocardial stress or ischemic injury causes cardiomyocyte necrosis or apoptosis, rapidly developing ventricular dysfunction, and sudden clinical deterioration requiring immediate intervention^[9]. In CHF, myocardial enlargement, stiffening due to fibrosis and calcification, and thickening (hypertrophy) ultimately worsen HF symptoms under excessive cardiac workload or abnormal stress^[10].

circRNAs play critical roles in both acute and chronic HF processes. In AHF, they can influence miRNA-protein interactions and regulate cardiomyocyte injury repair and inflammatory responses^[11]. They also function as “sponges” by binding miRNAs and thereby regulating the expression of miRNA targets, reducing cardiomyocyte apoptosis and even necrosis. For example, hsa-circRNAs modulate cardioprotective genes by competing for miR-133a binding and may help protect against AMI injury^[12]. In CHF, their roles are more complex: under chronic overload, circRNAs may regulate multiple signaling pathways involved in cardiomyocyte remodeling^[13], particularly those affecting hypertrophy, fibrosis, and apoptosis^[14]. Studies have shown that circRNAs interact with RBPs to affect cardiomyocyte metabolism and structure, thereby influencing cardiac remodeling pathways^[15]. Some circRNAs can interact with RBPs such as NF90 and FUS, enhancing cardiomyocyte adaptation to mechanical stress and slowing HF progression^[16].

At the molecular level, circRNAs regulate HF progression through several pathways. First, they inhibit miRNA activity through the “miRNA sponge” mechanism, relieving suppression of target genes involved in cardiac remodeling^[17]. They also regulate G protein-coupled receptor , PI3K/Akt, and MAPK/ERK signaling pathways to influence myocyte hypertrophy and apoptosis^[18]. Thus, circRNAs may be key regulators of HF onset and progression, as well as potential therapeutic targets requiring further validation. Advances in circRNA research have revealed their roles in regulating key target mRNAs at the transcriptional and translational levels, providing new perspectives on early HF diagnosis and treatment. Identifying specific circRNAs in patients with HF and combining these findings with gene-editing technology may offer avenues for personalized treatment. circRNAs therefore retain substantial research value and potential applications in HF. Their specific regulatory roles at different stages of HF development are summarized in Figure 1.

This review progresses from the basic biological characteristics of circRNAs to disease-related evidence and clinical applications. It first summarizes circRNA biogenesis and molecular functions, then discusses experimental evidence in AHF and CHF. Finally, it examines potential clinical applications, current methodological limitations, and future research directions.

This narrative review systematically collected relevant literature on the roles of circRNAs in AHF and CHF from PubMed, Web of Science, and Scopus. The search covered literature published through June 2025 and used the following keywords and their combinations: "circRNA", "circular RNA", "heart failure", "acute heart failure", "chronic heart failure", "cardiac hypertrophy", "cardiac remodeling", "fibrosis", "miRNA sponge", and "RNA-binding protein". Eligible publications included original research papers, review articles, and mechanistic studies published in English. Priority was given to literature providing evidence on circRNA biogenesis, molecular mechanisms, and experimental investigations of cardiovascular function. Additional relevant studies were identified by manually screening the reference lists of key publications. The included literature was systematically integrated and analyzed to summarize current knowledge of the biological functions and potential therapeutic value of circRNAs in heart failure.

circRNAs modulate gene expression and intracellular signal transduction through several mechanisms. These include acting as miRNA sponges through the ceRNA mechanism, interacting with proteins, and undergoing translation mediated by internal ribosome entry sites (IRESs) and m⁶A modification.

First, circRNAs may function as miRNA sponges that compete with miRNAs for binding to miRNA response elements (MREs), inhibiting miRNA binding to target mRNAs and relieving suppression of downstream gene expression. This is defined as the ceRNA mechanism^[30]. For example, CDR1as (ciRS-7), shown in Figure 1, contains more than 70 miR-7 binding sites and regulates the expression of miR-7 target genes^[31]. These findings support an important role for circRNAs in post-transcriptional regulation.

Second, circRNAs interact with RBPs to participate in diverse cellular processes^[32]. Some circRNAs act as “decoys” or “reservoirs” that regulate RBP activity, whereas RBPs contribute to circRNA formation and stability. For example, FUS and NF90/NF110 facilitate circRNA circularization by binding to intronic regions, while cir-CCND1 binds to HuR to upregulate CCND1 expression^[33]. Thus, circRNAs can function as effectors, receptors, or signaling intermediates and participate in post-transcriptional regulation.

In addition to their roles in post-transcriptional regulation and RNA interactions, some circRNAs have translational capacity. For example, circ-ZNF609 was the first circRNA confirmed to encode proteins in eukaryotic cells. Legnini et al. (Mol Cell, 2017) reported that circ-ZNF609 undergoes translation despite lacking a 5' cap and a 3'-poly(A) tail because it contains an IRES that enables cap-independent translation initiation, producing a 30 kDa protein. This translational activity increased markedly under heat shock stress and was associated with regulation of myocyte proliferation and differentiation^[34]. Despite the absence of a cap at the 5' end and a tail at the 3' end, internal IRES or N6-methyladenosine (m⁶A) sites can trigger circRNA translation^[35,36]. IRES-dependent translation allows circRNAs to initiate translation during stress or hypoxia, while m⁶A modification promotes circRNA translation initiation, with translational efficiency regulated by METTL3/4 methyltransferases and fat mass and obesity-associated protein (FTO) demethylase^[37]. These findings broaden the potential roles of circRNAs in cellular function and disease.

Overall, circRNAs regulate gene expression through three main mechanisms: miRNA sequestration, interactions with RBPs, and cap-independent translation. Their major molecular functions and regulatory mechanisms are summarized in Figure 3. These functions provide a molecular basis for circRNA involvement in cardiovascular disease onset and progression. Beyond their classical roles as miRNA sponges and protein-binding molecules, circRNAs can regulate the assembly or activity of protein-associated signaling complexes. For example, circIGF1R directly interacts with DDX5 and activates β-catenin signaling, promoting cardiomyocyte proliferation and cardiac repair after ischemic injury^[38]. circRNAs may also participate in epigenetic regulation through interactions with chromatin-modifying factors and RNA methylation-related mechanisms, including factors associated with m⁶A. These non-classical mechanisms broaden circRNA-mediated gene regulation and may play important roles in cardiac remodeling and heart failure progression.

circRNAs were previously considered little more than by-products of pre-mRNA splicing. In recent years, however, they have been identified as regulators of HF pathogenesis and progression. They play important roles in AHF and CHF by regulating miRNAs, participating in signal transduction, and maintaining cellular homeostasis, potentially providing new molecular targets for HF diagnosis and therapy. Ventricular dysfunction in HF is not caused by a single form of nonspecific cardiomyocyte injury but by heterogeneous pathological stimuli. Ischemic injury, such as acute myocardial infarction (MI), involves insufficient oxygen supply, mitochondrial dysfunction, oxidative stress, necrosis, apoptosis, inflammation, and scar formation. In contrast, pressure overload caused by hypertension or aortic valve stenosis primarily leads to concentric myocardial hypertrophy, activation of mechanical signaling pathways, fibroblast proliferation, extracellular matrix deposition, and subsequent deterioration of diastolic and systolic function. Volume overload, as occurs with valvular regurgitation or certain congenital heart diseases, more commonly manifests as eccentric ventricular remodeling, dilation, increased wall stress, and abnormalities in cytoskeletal and metabolic signaling. Although these conditions share downstream features, including inflammation, fibrosis, abnormal calcium handling, and ventricular remodeling, their upstream etiologies and temporal patterns of circRNA regulation may differ substantially. circRNA function must therefore be assessed in the context of the specific HF etiology and experimental model. Non-ischemic myocardial injury also contributes to HF onset and progression. Myocarditis caused by viral infection, autoimmunity, or other immune-mediated mechanisms can induce inflammatory cell infiltration, cytokine activation, cardiomyocyte necrosis or apoptosis, and subsequent fibrotic remodeling. Toxic cardiomyopathy, including anthracycline-induced injury, is typically associated with increased oxidative stress, mitochondrial damage, activation of DNA damage responses, impaired autophagy, and cardiomyocyte loss. These processes differ from ischemic injury and pressure overload but converge on common signaling pathways that lead to ventricular dysfunction. Current circRNA research in HF focuses mainly on ischemic and pressure overload models; studies of myocarditis and drug-induced cardiomyopathy remain relatively scarce. Future research should determine whether specific circRNA expression profiles distinguish ischemia, pressure overload, volume overload, inflammation, and toxic cardiomyopathy. Although many circRNAs have been associated with HF, they differ substantially in biological function, experimental evidence, and clinical translational potential. Table 1 summarizes representative circRNAs associated with AHF and CHF according to their molecular targets, experimental models, regulatory mechanisms, and functional effects.

Recent advances in sequencing technologies have substantially facilitated circRNA research. Long-read sequencing and full-length single-cell RNA sequencing have improved the characterization of full-length circRNA isoforms and their cell-type-specific expression patterns^[52-54]. However, circRNA identification and quantification remain technically challenging because most approaches rely on detecting back-splice junctions (BSJs), and different computational pipelines and algorithms may generate inconsistent candidate sets^[52-54]. Standardized sequencing protocols and bioinformatic workflows are therefore essential to improve reproducibility and comparability across studies.

Some circRNAs exhibit tissue-, cell-type-, and disease-stage-specific expression patterns^[55,56]. These characteristics may facilitate the development of circRNA-based molecular signatures, although their diagnostic and therapeutic utility requires further validation^[57].

The clinical feasibility of circRNA-based therapies also depends on delivery, safety, and regulatory considerations. Delivery platforms, including lipid nanoparticles, viral vectors, polymeric nanoparticles, and engineered extracellular vesicles or exosomes, need to achieve efficient, tissue-selective delivery while minimizing systemic exposure and unintended effects^[58,59]. Potential immunogenicity and inflammatory responses require careful evaluation, particularly for exogenously produced circRNAs and repeated administration^[58,59]. Moreover, the pharmacokinetic and pharmacodynamic properties of circRNA therapeutics, including biodistribution, intracellular persistence, degradation kinetics, and dose-response relationships, remain incompletely characterized^[59,60]. Clinical translation will also require scalable production, rigorous quality control, systematic safety assessment, and validation in appropriate preclinical and clinical studies^[59,60].

Despite their potential, most evidence on circRNAs in heart failure remains preclinical, with functional studies predominantly based on cultured cells and animal models and limited validation in large human cohorts^{[50,51,61-63]}. Tissue- and cell-type-specific expression may further complicate target selection and therapeutic delivery^[49,50]. Differences in study design, detection platforms, quantification methods, and bioinformatic pipelines also limit cross-study comparability and reproducibility^[52-54]. Multicenter human studies, independent validation cohorts, and standardized analytical procedures are therefore needed before circRNA-based biomarkers or therapies can be incorporated into routine clinical management^{[50,51,58,59,61-63]}.

This review discusses the complex roles of circRNAs in acute and chronic HF and their mechanisms of cardiac regulation. Further analysis of circRNA regulatory networks involved in myocardial stress responses, calcium homeostasis, energy metabolism, and fibrosis is needed to clarify their roles in these processes. Future studies should examine dynamic circRNA expression across pathological stages, tissue specificity, and interactions with signaling pathways, using integrative technologies such as high-throughput omics and single-cell sequencing to construct circRNA regulatory networks. The high stability, strong evolutionary conservation, and tissue-specific expression of circRNAs suggest potential applications in HF diagnosis. Future research should explore shared circRNA regulatory mechanisms across pathological states and systematically evaluate their value as diagnostic biomarkers and potential therapeutic targets. These applications require validation through further basic research and large-scale clinical studies. Detailed characterization of circRNA-related molecular networks and their interactions may deepen understanding of HF pathogenesis and support precision medicine strategies based on molecular classification and individualized intervention. Functional characterization is also needed to investigate circRNA-related multi-disease molecular patterns. Developing circRNA-related gene-editing or RNA-targeting drug strategies and improving in vivo delivery efficiency or circRNA-specific expression may provide further evidence to support precision medicine, early diagnosis, and targeted HF therapies. Future studies should also evaluate circulating circRNAs and exosome-derived circRNAs as minimally invasive biomarkers for HF diagnosis and prognostic assessment. Integrating advanced sequencing, bioinformatic tools, gene-editing platforms, and RNA delivery systems is expected to accelerate clinical translation and support precise diagnosis, risk stratification, and individualized HF treatment.

circRNAs are increasingly recognized as key regulators of the onset and progression of HF. They play important roles in cardiac pathophysiology by linking post-transcriptional regulation, protein interaction networks, and stress-response signaling pathways. Current evidence indicates that circRNAs are involved in multiple pathological processes closely associated with HF, including cardiomyocyte injury, abnormal calcium handling, apoptosis, myocardial hypertrophy, fibrosis, inflammation, and ventricular remodeling.

Mechanistically, circRNAs not only act as miRNA sponges but also bind proteins, potentially serve as protein scaffolds, and regulate RNA modifications and translation-related processes. These diverse biological functions suggest that circRNAs may participate in both adaptive and maladaptive cardiac responses in AHF and CHF.

Despite substantial progress in recent years, clinical translation remains at an early stage. Most findings on circRNA function derive from cell-based experiments and animal models, and their relevance to human heart failure requires further validation. Unresolved issues include stage-specific effects, cell-type heterogeneity, difficulties in detecting low-abundance transcripts, inconsistent bioinformatics workflows, small clinical sample sizes, and insufficient standardization of circRNA quantification. Future research should therefore prioritize rigorous validation of post-splicing sites, multicenter human studies, integration of single-cell sequencing and spatial transcriptomics, and standardized data analysis workflows to improve reliability and reproducibility. Given their high stability and ease of detection in body fluids, circRNAs are most likely to find initial clinical applications as biomarkers for heart failure diagnosis, risk stratification, and disease monitoring, particularly as circulating circRNAs or exosome-derived circRNAs. Therapeutic strategies based on circRNAs, including circRNA silencing, functional replacement, engineered exosome delivery, nanoparticle-mediated delivery, and CRISPR/Cas-mediated precise regulation, also show promise. However, these approaches require further optimization and validation of targeting specificity, safety, delivery efficiency, and long-term efficacy.

Overall, circRNAs provide an important framework for understanding the molecular mechanisms of heart failure. However, most available evidence derives from preclinical models of myocardial infarction, pressure overload, and cardiac remodeling rather than large-scale clinical studies. Evidence for HFpEF and heart failure with HFmrEF also remains limited. Current findings should therefore be interpreted cautiously and should not be generalized to all cases of acute and chronic heart failure without further clinical validation.