Unifying and unique roles of non-coding RNA biomarkers in liver and heart fibrosis Authors: Veronika Boichenko, Svetlana Frolova, Christine Voellenkle, Germana Zaccagnini, Paolo Ferrero, Pier Paolo Bassareo, Giovanni Battista Luciani, Julia Gorelik, Fabio Martelli Published: 2026 ABSTRACT AbstractLiver and cardiac fibrosis are critical drivers of chronic organ dysfunction, characterised by fibroblast activation, excessive extracellular matrix deposition, and inflammation. While these fibrotic processes are initiated by distinct triggers and monitored by organ-specific biomarkers, emerging evidence reveals both shared and unique regulatory networks, particularly involving non-coding RNAs (ncRNAs). This review focuses on the roles of microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs), such as miR-21, miR-22, miR-29, miR-34a, miR-122, miR-133a, miR-210, miR-214, H19, MALAT1, MEG3, NEAT1, and circHIPK3, in modulating key fibrotic pathways including TGF-β/SMAD signalling, oxidative stress, and extracellular matrix remodelling in both hepatic and cardiac tissues. Some ncRNAs exhibit convergent regulatory functions across both organs, while others demonstrate divergent or context-dependent effects within and between tissues. Furthermore, circulating ncRNAs, potentially transferred via extracellular vesicles (EVs), hold potential as minimally invasive biomarkers for early detection of combined cardio-hepatic fibrosis, as exemplified in conditions like Fontan circulation, though extensive validation in large, prospective cohorts is essential. This article summarises established clinical biomarkers, discusses the ncRNA-mediated mechanisms operating in liver and heart fibrosis, distinguishing shared, divergent, and mixed evidence, and explores their translational potential in diagnostics and experimental therapeutics. Finally, it addresses remaining challenges and outlines future directions, emphasising multi-omic integration, longitudinal ncRNA profiling, and mechanistic validation to advance precision management of fibrotic diseases.Keywords: Liver fibrosis, Cardiac fibrosis, Circulating biomarkers, Non-coding RNAs, miRNAs, lncRNAs Introduction Cardiac and liver fibrosis connection in disease development Fibrosis represents a conserved pathological response to chronic tissue injury and is characterised by persistent fibroblast activation, excessive accumulation of collagen-rich extracellular matrix (ECM), and unresolved inflammation [1]. In both the heart and the liver, sustained fibrotic remodelling disrupts normal tissue architecture, ultimately leading to progressive organ dysfunction [1], [2], [3]. Importantly, accumulating evidence supports the existence of a cardio-hepatic axis, in which pathological processes in one organ can influence fibrogenesis in the other through systemic haemodynamic, metabolic, and inflammatory signals. This inter-organ axis is further supported by the shared core regulatory programmes, including transforming growth factor-β (TGF-β) signalling, myofibroblast differentiation, dysregulated ECM turnover, and chronic immune cell engagement [1], [2], [3]. Nevertheless, despite substantial mechanistic overlap, fibrosis also exhibits tissue-specific features shaped by distinct cellular composition, vascular structure, and regenerative capacity. For example, cardiac fibrosis is closely associated with cardiomyocyte injury and impaired contractility, leading to adverse ventricular remodelling [2]. In contrast, liver fibrosis arises from diverse metabolic, toxic, and inflammatory insults and retains a greater potential for reversibility [3]. The present review focuses on non-coding RNAs (ncRNAs) as plausible mechanistic regulators of both processes and discusses their potential utility for monitoring fibrosis-associated diseases. Clinically, cardiac and hepatic fibrosis may coexist or develop sequentially. Cardiovascular disease (CVD) can promote liver injury through venous congestion and systemic inflammation, whereas chronic liver disease (CLD) contributes to cardiac remodelling via metabolic dysregulation, altered haemodynamics, and the release of circulating profibrotic mediators [4], [5], [6]. Cardiac fibrosis represents a hallmark of most forms of CVD [2], while liver fibrosis underlies the progression of CLDs, including viral hepatitis, alcohol-related liver disease, non-alcoholic steatohepatitis (NASH), cholestatic disorders, and congestive hepatopathy [3]. Collectively, fibrotic diseases of the heart and liver represent a major contributor to global morbidity and mortality [7]. By examining cardiac and liver fibrosis in parallel, this review aims to propose ncRNAs, which may serve as candidates for future large-scale studies. The ultimate goal is to enable shared candidate diagnostic markers and mechanistically guided risk stratification, thereby supporting earlier detection and more precise intervention across fibrotic diseases, contingent upon overcoming the current research limitations [1], [4], [6]. Fibrosis and remodelling of the heart Pathological cardiac remodelling is a major driver of progression to heart failure (HF) [8]. It involves structural and functional changes of the heart, including cardiomyocytes (CMs) hypertrophy, fibrosis, chamber dilation, and altered electrical conduction [8]. Among these features, fibrosis plays a central role by disrupting myocardial architecture and contributing to both systolic and diastolic dysfunction [2]. Cardiac fibroblasts (CFs) are the main regulators of ECM homeostasis and, upon activation, differentiate into myofibroblasts that excessively deposit ECM proteins, including collagens I, III, and VI, thereby promoting tissue stiffening [2]. This activation is driven by a complex network of pro-fibrotic and inflammatory signalling, including TGF-β and interleukin-11 (IL-11), which position CFs as key contributors to pathological remodelling and the primary therapeutic targets in CVD [2], [9], [10]. Accordingly, cardiac fibrosis is widely recognised as a key process through which diverse insults, including ischemic, inflammatory, mechanical, and genetic injury, converge to promote ECM accumulation and ventricular dysfunction [2]. These mechanistic insights have shaped contemporary approaches to CVD assessment and management, guiding the development of circulating and imaging-based biomarkers, such as natriuretic peptides and cardiac MRI-based fibrosis measures. In parallel, they have supported the use of disease-modifying therapies including β-blockers and renin-angiotensin-aldosterone system (RAAS) inhibitors, particularly in HF [11]. HF is commonly classified according to left ventricular (LV) ejection fraction into HF with preserved ejection fraction (HFpEF) and HF with reduced ejection fraction (HFrEF) [9]. HFpEF is characterised by LV hypertrophy and wall thickening with preserved systolic function, whereas HFrEF involves ventricular thinning, impaired contractility, and reduced cardiac output [12]. In HFrEF, cardiac fibrosis predominantly arises from cardiomyocyte loss that leads to replacement fibrosis [13]. In HFpEF, excessive collagen deposition coupled with a relative reduction in the more compliant collagen type III contributes to increased myocardial stiffness [13]. Although myofibroblast activation occurs in both conditions, differences in ECM regulation and pro-fibrotic signalling may contribute to distinct patterns of fibrotic remodelling. For example, HFrEF has been associated with higher MMP-9/TIMP-1 ratios compared to HFpEF [14], while higher TGF-β1 concentrations have been reported in HFpEF patients than HFrEF patients [15]. Fibrosis and remodelling of the liver Liver fibrosis is a dynamic wound-healing response to chronic liver injury, arising from diverse aetiologies such as viral infection, excessive alcohol consumption, metabolic disorders, or toxic exposure, that drives ECM accumulation that disrupts liver structure and function [16]. Following acute liver injury, parenchymal cells undergo efficient regeneration to replace necrotic and apoptotic cells and restore tissue integrity. In contrast, during persistent or repetitive injury, regenerative capacity progressively declines, and excessive ECM deposition increasingly replaces functional hepatocyte mass, leading to the development of hepatic fibrosis [17]. As chronic injury advances, multiple hepatic cell populations, including both parenchymal and non-parenchymal cells, undergo distinct phenotypic and functional changes that collectively drive fibrogenesis [18]. Kupffer cells and resident liver macrophages are activated, producing cytokines, chemokines, and key mediators such as TGF-β and platelet-derived growth factor (PDGF). These signals not only initiate hepatic stellate cell (HSC) activation but also sustain and amplify fibrotic responses within the injured liver [19]. HSC are central players of fibrogenesis: while quiescent in a healthy liver, they differentiate into myofibroblast-like cells upon injury, acquiring the capacity to secrete ECM components, and drive scar formation [20]. Importantly, liver fibrosis is no longer viewed as a strictly irreversible process [16]. Both experimental and clinical evidence have demonstrated that fibrosis can regress following removal of the underlying injurious stimulus, particularly at earlier stages. Fibrosis regression involves coordinated mechanisms including HSC inactivation or apoptosis, resolution of inflammation, and remodelling of excess ECM [20]. Nevertheless, persistent injury promotes progression to advanced fibrosis and cirrhosis [21]. These conditions are marked by nodular regeneration that disrupts normal liver structure, portal hypertension due to impaired blood flow through the fibrotic liver and a higher risk of hepatocellular carcinoma (HCC) [21]. Although fibrotic histological features are shared across liver diseases, the upstream drivers and spatial patterns of ECM deposition vary depending on disease context, including metabolic, inflammatory, or haemodynamic stressors [22]. Cardiac disease as a driver of liver injury The association between CVD, particularly HF, and liver disease is well established [5]. HFpEF is frequently associated with systemic inflammation, microvascular complications, overweight or obesity, type 2 diabetes (T2DM) and metabolic dysfunction-associated steatotic liver disease (MASLD) [23]. Mechanistically, cardiometabolic phenotype of HFpEF, at least partially, is linked to myocardial steatosis, accumulation of lipid in heart, and systemic inflammation, making this organ damage similar to MASLD [24], [25]. Key features of HFrEF such as changed sympathetic nervous system activity, secondary hyperaldosteronism, and fluid retention overlap with cirrhosis [24]. In HFpEF, LV diastolic dysfunction increases left atrial and pulmonary pressures, promoting pulmonary hypertension and RV dysfunction in about 20–35% of patients [26]. This can lead to cardiac hepatopathy, often coexisting with MASLD and worsening disease severity cardiac hepatopathy [23]. In HFrEF, reduced LV contractility lowers cardiac output and increases left-sided filling pressures, leading to secondary pulmonary hypertension [27]. Progressive RV decompensation may ultimately culminate in right-sided HF and cardiac hepatopathy [28]. In HF with right ventricular dysfunction, impaired flow into the pulmonary circulation leads to venous backflow and elevated pressures within the hepatic veins and sinusoids, resulting in tissue hypoxia and progressive liver fibrosis [5], [29], [30], [31]. This congestive pathway to fibrosis and eventual cirrhosis also occurs in other conditions associated with chronically elevated central venous pressure [5], [32], [33]. These include constrictive pericarditis and Fontan circulation [5], [32], [33]. Fontan circulation is a surgically created haemodynamic state used to treat univentricular heart disease, in which systemic venous blood is routed directly to the pulmonary arteries without the support of a sub-pulmonary pumping chamber [34]. As a result, chronic, non-pulsatile venous congestion becomes an essential haemodynamic feature rather than a compensatory abnormality [5], [34]. This unique physiology leads to early and nearly universal development of Fontan-associated liver disease (FALD), which is distinct from conventional HF-related congestive hepatopathy due to persistent elevation of central venous pressure transmitted directly to the hepatic veins and sinusoids [30], [32], [34] (Fig. 1). Schematic of liver haemodynamics in biventricular vs single-ventricle (Fontan) circulation. The upper part of Panel A depicts normal biventricular circulation, in which blood from the bowel passes through the liver and is pumped to the lungs by a sub-pulmonary ventricle, compared with Fontan circulation (lower part of Panel A), where venous blood flows passively to the pulmonary arteries in the absence of a pumping chamber. Red arrows indicate oxygenated blood, and blue arrows deoxygenated blood. Panel B illustrates pressure profiles along this pathway: in biventricular circulation, pressure decreases from the bowel to the liver and rises sharply across the right ventricle (solid line), whereas in Fontan circulation, pressure declines more linearly from the bowel to the lungs (dashed line), resulting in chronically elevated portal and sinusoidal pressures [30], [34]. These activate pro-fibrotic molecular pathways, hypoxia-responsive pathways, and mechanotransduction cascades, driving hepatic fibrogenesis. In parallel, Fontan patients are also prone to the development of myocardial fibrosis. Cardiac MRI-detected fibrosis has been shown to correlate with diastolic dysfunction, increased liver stiffness, and elevated circulating fibrosis biomarkers, supporting the concept of coupled fibrotic remodelling of the heart and liver driven by shared haemodynamic and molecular stressors [35]. With improving survival and growing adult Fontan population, interest in FALD has increased, highlighting the importance of early screening and surveillance, especially in the context of transplantation planning [36]. Although FALD may clinically resemble toxic or infectious liver disease, it differs in its pathophysiology (Table 1). Table 1Clinical profile of cardiac-related and non-cardiac liver disease.Disease featureCardiac-related liver diseaseOther aetiologies (toxic, infection, etc.)References (cardiac-related)References (other aetiologies)Inflammationuncommoncommon[31][19]Sinusoidal congestioncommonuncommon[30][37]Fibrosiscommoncommon[30][3]Main locus of pathologycentrolobularportal spaces[31][3]Portal hypertensioncommon, post-hepaticcommon, intra-hepatic[32][38]Esophageal varicesuncommoncommon[29][38] Clinical profile of cardiac-related and non-cardiac liver disease. Emerging evidence suggests that liver injury extends beyond haemodynamic congestion to involve molecular crosstalk between the heart and liver [5], [39]. Hernandez-Anzaldo et al. proposed an endocrine function of the heart that influences liver function [40]. In Mmp2−/− mice, the heart releases phospholipase (sPLA2), which triggers liver inflammation and abnormal lipid metabolism [40]. Blocking cardiac sPLA2 reversed these liver abnormalities, revealing a previously unrecognised heart-liver signalling axis [40]. Despite clinical recognition of cardio-hepatic interactions, the molecular mediators underlying this inter-organ communication remain poorly defined [29]. Liver disease as a driver of cardiac injury Cardiac complications in liver disease arise from a cascade of haemodynamic, inflammatory, metabolic, and neurohormoral disturbances [4], [6], [41], [42]. CLD, including cirrhosis of viral, alcohol-related, or metabolic origin, causes profound circulatory changes, driven by inflammation and endothelial dysfunction. Increased vasodilator production widens blood vessels, lowering resistance to blood flow [41]. The resulting hyperdynamic circulation is marked by elevated cardiac output, though effective tissue perfusion remains reduced [41]. In response, neurohormonal systems, such as the RAAS and the sympathetic nervous system, are chronically activated to maintain perfusion [41]. In parallel, vascular congestion within the splanchnic bed and subsequent development of intestinal edema, lead to translocation of gut-derived bacterial products into the circulation. This process drives systemic inflammation, oxidative stress, and endothelial dysfunction [41]. Circulating cytokines (such as IL-1β and IL-6) directly affect the myocardium, promoting CF activation and ion channel dysfunction, thereby contributing to cardiac fibrosis and arrhythmia [4], [41]. These alterations also disrupt normal ventricular filling and impair myocardial contractile reserve, predisposing to cirrhotic cardiomyopathy and HF [41]. Metabolic liver diseases, such as non-alcoholic fatty liver disease (NAFLD), its progressive form NASH, and metabolic-associated fatty liver disease (MAFLD), also increase cardiovascular risk. These conditions are associated with insulin resistance and altered lipid metabolism, which impair CM energetics [4], [42]. Comorbidities such as diabetes and obesity exacerbate both liver disease and cardiac complications [41], [43]. Although cardio-hepatic crosstalk is well recognised clinically, molecular mediators responsible for this communication remain largely undefined [29]. Few circulating proteins have been proposed, including liver-derived protein coagulation factor XI, which protects against diastolic dysfunction [44], and hepatic fibroblast growth factor 21 (FGF21), which mitigates myocardial infarction (MI) [45]. Soluble leukemia inhibitory factor receptor (LIFR) has been described as a hepatocyte-derived mediator that promotes CFs migration [46]. Together, these studies support the existence of liver-derived mediators contributing to cardiac remodelling. Overall, CLD actively drives cardiac injury through haemodynamic overload, systemic inflammation, metabolic dysfunction, and liver-derived factors. Understanding these liver-heart interactions is critical for risk stratification, early detection, and the development of therapies to mitigate cardiac complications in patients with CLD. Current biomarkers of cardiac fibrosis In clinical practice, the most commonly measured circulating cardiac biomarkers are B-type natriuretic peptide (BNP), its N-terminal fragment (NT-proBNP), and cardiac troponins (cTns) [47]. Although highly cardiac-specific, BNP and NT-proBNP primarily reflect myocardial wall stress and ventricular stretch rather than fibrosis [11], [47], [48]. Moreover, BNP and NT-proBNP also reflect haemodynamic status [49]. Consequently, their levels are elevated across diverse CVD [11], [47], [48]. Nevertheless, higher BNP/NT-proBNP levels stratify prognosis in patients with and without clinically overt HF, supporting their value for risk assessment [50]. High-sensitivity cardiac troponins (cTns) are mainly used to detect MI; however, elevated levels can also be seen in HF [47], [51]. Higher circulating cTnT levels independently predict adverse outcomes in HF patients, even in the absence of acute MI [52]. Nevertheless, cTns lack of specificity limits their diagnostic utility in HF [47], [51]. Notably, both NT-proBNP and cTnT are linked to RAAS activation, reflecting neurohormonal stress and subclinical myocardial injury [53]. Galectin-3 (Gal-3) has emerged as an additional marker of cardiac fibrosis [54]. Like BNP and NT-proBNP, Gal-3 is associated with morbidity and mortality [55], but it reflects intrinsic fibrotic remodelling rather than haemodynamic stress [54]. Despite these findings, no blood-based biomarker of cardiac fibrosis has yet demonstrated sufficient reliability for routine clinical use [56]. These remain challenging to identify due to their lack of cardiac specificity, overlap with systemic inflammation, and the heterogeneous nature of cardiac remodelling. As a result, many available markers reflect general tissue injury or systemic processes rather than localised cardiac fibrotic activity [57]. Consequently, current assessment methods of cardiac fibrosis rely largely on imaging techniques such as cardiac MRI with late gadolinium enhancement and T1 mapping [57]. While useful, these approaches fail to capture the dynamic molecular processes driving fibrosis [57]. Monitoring fibrosis at this level will require the development of novel, mechanism-specific biomarkers, such as ncRNAs [57]. Current biomarkers of liver fibrosis Liver fibrosis can be assessed using established biomarker panels, including Enhanced Liver Fibrosis (ELF), N-terminal propeptide of type III collagen (PRO-C3), and cytokeratin-18 (CK-18) [58]. Currently used liver fibrosis and cardiac biomarkers are summarised in Table 2. Table 2Comparison of circulating biomarkers associated with cardiac and hepatic fibrosis.Cardiac fibrosisLiver fibrosisRef.(cardiac fibrosis)Ref.(liver fibrosis)BiomarkerBNP, cTns, Gal-3ELF, PRO-C3, CK-18[47][58]Aetiology specificitypoorpoor[47][59]RAAS activation associationcommonuncommon[53][60]Correlation with haemodynamicsgoodpoor[49][61], [62], [63]Correlation with local fibrogenesisinconsistentinconsistent[11], [47], [48][64], [65]Association with disease prognosisvalidated in large cohortsinsufficient validation[50], [52], [55][66], [67] Comparison of circulating biomarkers associated with cardiac and hepatic fibrosis. As highlighted in Table 2, the currently available circulating biomarkers of liver fibrosis reflect ECM turnover and hepatocellular injury. The comparatively stronger establishment of circulating liver fibrosis biomarkers may reflect fundamental aspects of liver biology. As a major source of plasma- and serum-associated proteins [68], the liver could, during fibrotic injury, more readily influence the circulating milieu, for example, by increasing the release of enzymes and ECM-derived fragments [58]. However, interpretation of these biomarkers should remain cautious. Their performance is limited by poor aetiology specificity [59] and inconsistent correlation with local fibrogenesis [64], [65]. Moreover, associations between liver fibrosis biomarkers and RAAS activation have not been established [60]. In addition, liver fibrosis biomarkers show limited relationships with systemic or portal haemodynamics [61], [62], [63] and have not been consistently linked to liver disease prognosis [66], [67]. While ELF may partially reflect portal haemodynamics, PRO-C3 and CK-18 predominantly capture fibrogenic activity that appears independent of circulatory congestion [61], [62], [63]. Congestive hepatopathy represents a distinct pathogenesis compared with viral, metabolic, or cholestatic liver disease, resulting in poor performance of most non-invasive fibrosis markers in this setting [69]. Although the Model for End-Stage Liver Disease (MELD) [70] and its adaptation for anticoagulated patients (MELD-XI) are widely used, MELD-XI shows only modest correlation with histological fibrosis [71], [72]. Likewise, FIB-4 and APRI, developed for hepatitis C, show limited reliability in cardiac-induced liver disease, as reinforced by a recent meta-analysis [73], [74], [75]. Overall, current circulating fibrosis biomarkers lack specificity and reliability, especially in congestive hepatopathy. These limitations highlight the need for integrated biomarkers that better capture shared cardio-hepatic fibrotic mechanisms. Overview and selection of non-coding RNAs The human transcriptome is dominated by ncRNA transcripts, whereas protein-coding messenger RNAs (mRNAs) account for less than 3% of all genomic transcription [76]. Advances in high-throughput sequencing and complementary technologies have facilitated the discovery of a wide diversity of ncRNA species with varied structural and size characteristics [77], [78]. Collectively, the ncRNA family comprises a vast collection of transcripts that lack the ability to be translated into proteins [79]. Based on nucleotide (nt) length, ncRNAs are categorised into small (<200 nt) and long (>200 nt) ncRNAs [79]. Among non-coding transcripts, microRNAs (miRNAs), long ncRNAs (lncRNAs) and circular RNAs (circRNAs) have been recognised as essential regulators in various biological processes and have been associated with various human diseases [9], [80], [81], [82]. Beyond their intracellular functions, ncRNAs are stable in the circulation and can be readily detected in biological fluids, including plasma [83], [84], serum [85], [86], and whole blood [87] across diverse pathological conditions. Altered circulating ncRNA profiles in heart and liver fibrosis may serve as candidate biomarkers for disease detection, phenotyping, and mechanistic insight into fibrotic pathways, though their clinical application is still in early stages of investigation and requires rigorous validation [9], [88], [89], [90], [91]. To identify lncRNAs and circRNAs implicated in both CLD and CVD fibrotic disease, manual literature searches were conducted in PubMed for each ncRNA class separately, all restricted to January 2015 – November 2025. For circRNAs, the queries 1) [("circRNA" OR "circular RNA") AND "liver fibrosis"] and 2) [("circRNA" OR "circular RNA") AND "cardiac fibrosis"] were used, while for lncRNAs: 1) [("long non-coding RNA" OR "lncRNA") AND "liver fibrosis"] and 2) [("long non-coding RNA" OR "lncRNA") AND "cardiac fibrosis"]. Review articles, systematic reviews, meta-analyses, editorials, and preprints were excluded from both searches, with cancer-related articles additionally excluded for lncRNAs, yielding 38 and 26 original research articles for circRNA in liver and cardiac fibrosis, respectively, and 110 and 136 for lncRNA in liver and cardiac fibrosis, respectively. circRNA and lncRNA names were extracted from titles and abstracts and cross-referenced between the two datasets. This identified circHIPK3 as the only circRNA shared by both conditions. For lncRNAs, 16 shared candidates were found, of which four (H19, MEG3, NEAT1, and MALAT1), were selected for narrative review. Their presence in circulating biofluids was also confirmed through targeted PubMed searches using each ncRNA name together with the terms “cardiac”, “heart”, “liver”, or “hepatic”, and “plasma”, “serum”, “blood”, “circulating”, or” biomarker” (Fig. 2). Workflow for ncRNA selection in hepatic and cardiac fibrosis. PubMed literature was screened for miRNAs, lncRNAs, and circRNAs associated with liver and cardiac fibrosis, followed by overlap analysis and assessment of evidence for candidate ncRNAs modulated in circulation. Given the considerably larger volume of miRNA-related literature, a computational screening pipeline was developed with the NCBI E-utilities via Biopython, applying the same publication filters. miRNAs were extracted from PubMed abstracts using a standardised regular expression pattern (miR-\d+), and only those appearing in both liver and cardiac fibrosis datasets were retained. This yielded 589 articles for liver fibrosis (from which 392 unique miRNAs were extracted) and 390 articles for cardiac fibrosis (from which 215 unique miRNAs were extracted). The common miRNAs identified in both conditions (overlap: 120 miRNAs) were then assessed for their candidate biomarker potential. A subsequent check revealed that at least 100 miRNAs were reported as being modulated in circulation (Fig. 2). Considering the high number of potential candidates, eight miRNAs were selected for narrative review based on manual literature searches and prior expertise: miR-21, miR-22, miR-29, miR-34a, miR-133a, miR-122, miR-210 and miR-214. It should be noted, however, that many of the studies fulfilling these selection criteria, particularly those involving human cohorts, frequently exhibit methodological limitations, including small sample sizes, absence of independent validation cohorts, and incomplete adjustment for confounding variables, which collectively temper the immediate translational relevance of the reported findings [1], [4], [6]. miRNAs miRNAs are ∼22 nt long RNAs that regulate target gene expression through sequence-specific recognition via the RNA-induced silencing complex (RISC) [92], [93]. This leads to mRNA degradation, deadenylation, or translational repression [92], [93]. Because miRNA-mRNA networks are many-to-many (one miRNA targets multiple transcripts, and individual genes are regulated by multiple miRNAs), miRNAs influence diverse cellular pathways [94], [95], [96]. Owing to their stability and detectability in circulation, altered miRNA levels may provide diagnostic insights and indicate disease progression [9], [97], [98], [99]. Some technical limitations affecting their biomarker utility are discussed in Section 2.4. Table 3 lists miRNAs implicated in liver- and heart fibrosis-associated diseases, highlighting their molecular mechanisms and regulatory roles in fibrotic processes. Table 3Key miRNAs and their mechanisms of action in fibrosis of the liver and heart.OrganExperimental systemRole in fibrosisMechanismRef.microRNA-21LiverLX-2promiR-21→↓PTEN→↑PI3K/AKT[100]rHSC, in vivo micepromiR-21→↓SMAD7→↑TGF-β[101], [102]rHSCpromiR-21→↓SPRY2 →↑ERK1[103]HeartmCFspromiR-21 →↑TGF-β[104]rCFspromiR-21→↓CADM1→↑STAT3[105]rCFs, in vivo ratpromiR-21→↓Jagged1→↓Notch[106]mCFs, in vivo micepromiR-21→↓Spry1→↑ERK-MAPK[107]microRNA-22LivermHSC, in vivo miceantiNeat1→↓miR-22-3p→↑cyth3[108]LX-2, in vivo ratsantimiR-22-3p→↓AKT3[109]in vivo micepromiR-22→↓BMP7[110]HeartrCFspromiR-22→↓Cav3→↑PKCε[111]mCFsantimiR-22-3p→↓PTAFR[112]hCFs, in vivo micepromiR-22→↓OGN[113]rCFs, in vivo miceantimiR-22→↓TGFβRI[114]rCFs, in vivo ratsantimiR-22→↓PTEN/Akt/mTOR[115]microRNA-29 familyLiverin vivo miceantimiR-29a→↓MCJ, CDK6[116]LX-1, in vivo miceantimiR-29b→↓ PI3K/AKT[117]Heartin vivo micepromiR-29→↓ (GSK3B, CTNNBIP1) →↑Wnt[118]rCFs, in vivo ratsantiAMK→↓TGF-β1→↑miR-29 →↓CDK2[119]microRNA-34aLiverin vivo ratpropositive loop: p53→↑miR-34a→↓SIRT1→↑p53[120]HeartrCFs, in vivo ratsproTGF-β1→↑miR-34a→↓c-ski→↑TGF-β[121]microRNA-122LiverLX-2, in vivo miceantimiR-122→↓P4HA1[122]LX2, in vivo miceantiTGF-β→↓miR-122→↑FN1[123]in vivo miceantimiR-122a→↓KLF6→↓TGF-β1[124]HeartrCFs, in vivo ratspromiR-122-5p→↓ apelin→↓ACE2, GDF15[125]mCFs, in vivo miceantiapigenin→↑miR-122-5p→↓HIF-1α[126]hCFsantimiR-122→↓TGF-β[127]microRNA-133aLiverLX-2, mHSCantimiR-133a→↓Col1A1, Col5A3[128]LX-2antimiR-133a→↓Smad3[129]Heartin vivo ratsantimiR-133a-3p→↑SOX5[130]in vivo miceantimiR-133a→↓TGF-β1→ ↓ERK1/2[131]in vivo ratsantimiR-133a→↓CTGF and TGF-β1[132]microRNA-210HeartrCFantimiR-210-5p→↓TGFBR1[133]microRNA-214LiverLX-2, in vivo miceantimiR-214-3p→↓MECP2[134]rHSC, LX-2proTwist1→↑miR-214-3p→↓Sufu→↑Hedgehog[135]mHSC, LX-2antiTwist1→↑miR-214→↓CCN2[136]HeartmCFs, in vivo miceantimiR-214-3p→↓NLRC5[137]mMFs/ in vivo miceantimiR-214-3p→↓EZH1/2→↑PPAR-γ[138]mCFs, in vivo miceantimiR-214-3p→↓caspase-1→↓pyroptosis[139]in vivo micepromiR-214-3p→↓c-Ski[140]in vivo miceantimiR-214→↓Ncx1, CaMKII, CypD, BIM[141]rCFs, in vivo ratsantimiR-214 → ↓TGF-β1[142]rCFs, in vivo ratspromiR-214 → ↓ Mfn2 → ↑ERK1/2[143]CFs, cardiac fibroblasts; hCFs, human CFs; mCFs, mouse CFs; rCFs, rat CFs; LX-1, LX-2, human hepatic stellate cell line; mHSCs, mouse hepatic stellate cells; mMFs, mouse myofibroblasts; rHSCs, rat hepatic stellate cells. “anti” indicates anti-fibrotic; “pro” indicates pro-fibrotic. Key miRNAs and their mechanisms of action in fibrosis of the liver and heart. CFs, cardiac fibroblasts; hCFs, human CFs; mCFs, mouse CFs; rCFs, rat CFs; LX-1, LX-2, human hepatic stellate cell line; mHSCs, mouse hepatic stellate cells; mMFs, mouse myofibroblasts; rHSCs, rat hepatic stellate cells. “anti” indicates anti-fibrotic; “pro” indicates pro-fibrotic. lncRNAs lncRNAs are defined as transcripts longer than 200 nt that lack protein-coding potential [144]. Once considered transcriptional “noise”, lncRNAs are now recognised as potential regulators of gene expression and cellular processes [145]. lncRNAs employ multiple regulatory mechanisms. These include recruitment of chromatin-modifying complexes, scaffolding of protein assemblies, sequestration of transcription factors, and organisation of three-dimensional chromatin architecture, to precisely control gene expression across diverse cellular contexts [146], [147]. Functionally, lncRNAs can act as molecular signals directing epigenetic modifications (e.g., HOTAIR-mediated histone demethylation) [148], and as structural platforms organising ribonucleoprotein complexes (e.g., NEAT1 in paraspeckles [149]). In addition, lncRNAs can function as molecular decoys by interacting with proteins or other RNAs, preventing them from interacting with their targets [150]. A prominent regulatory role of lncRNAs is their function as competing endogenous RNAs (ceRNAs), whereby lncRNAs act as miRNA sponges, sequestering miRNAs and protecting target mRNAs from miRNA-mediated silencing [151]. The combinatorial nature of these mechanisms highlights lncRNAs as sophisticated regulatory molecules capable of integrating cellular signals and modulating gene expression networks with remarkable specificity and flexibility [147]. Some lncRNAs are regulators of fibrosis with pro-fibrotic [152], [153], [154] and anti-fibrotic functions [154], [155]. Some lncRNAs have also been found to be dysregulated in biofluids of patients with hepatic and cardiac fibrosis, supporting their potential as circulating biomarkers, although current evidence remains preliminary [156], [157], [158], [159]. lncRNAs implicated in liver and heart fibrosis, together with their mechanisms, are summarised in Table 4. Table 4Key lncRNAs and circRNAs: their mechanisms of action in fibrosis of the liver and heart.OrganExperimental systemRole in fibrosisMechanismRef.H19LiverLX-2, in vivo miceproH19→↓miR-148a →↑TGFbRI[160]mHSC, in vivo miceproH19→↑Wnt/β, apoptosis[161]HSC, in vivo ratsproDNMT1→↓H19→ ↑ERK1/2[162]HeartrCFsproH19→↓miR-29a/b→↑TGF-β[163]rCFs, in vivo ratsproH19→↓DUSP5→↑ERK1/2[164]mCFs, in vivo miceproH19→↓miR-455→↑CTGF[165]MEG3LivermHSC, in vivo miceantiMEG3→↓SMO[166]LX2, in vivo miceantiMEG3 →↓NLRC5[167]LX2, in vivo miceantiMEG3→↑p53[168]LX2antiMEG3→↓miR-145→↑PPARγ[169]in vivo miceantiMEG3→↑SIRT6[170]Heartin vivo miceproMEG3→↓miR-129→↑ATG14[171]in vivo miceproMEG3→↓miR-223→↑NLRP3[172]MALAT1LiverLX2proMALAT1→↓miR-181a→↑TLR4/NF-κB[173]mHSC, in vivo miceproMALAT1→↓miR-101b→↑Rac1[174]LX-2proMALAT1→↑CXCL5[175]HeartmCFs, in vivo miceproMALAT1→↓miR-145→↑TGF-β1[176]mCFs, in vivo miceproMALAT1→↓miR-141→↑NLRP3, ↑TGF-β1[177]NEAT1LivermHSC, in vivo miceproNEAT1→↓miR-122→↑KLF6[178]LX-2, in vivo miceproNEAT1→↓miR-139-5p→↑β-catenin→↑ TGF-β1[179]mHSC, in vivo miceproNeat1→↓miR-148a-3p/miR-22-3p→↑Cyth3[108]mHSC, in vivo miceproNEAT1→↓miR-129-5p→↑PEG3[180]mHSC, in vivo miceproNEAT1→↓miR-29b→↑Atg9a[181]HeartmCFs, in vivo miceproNEAT1→↓Smad7[182]circHIPK3Liverin vivo miceprocircHIPK3→↓IGF2BP2/PPAR-γ[183]HeartmCFs, in vivo miceprocircHIPK3→↓miR-152→↑TGF-β2[184]CFs, cardiac fibroblasts; mCFs, mouse CFs; rCFs, rat CFs; HSC, hepatic stellate cells; mHSC, mouse HSC; LX-2, human hepatic stellate cell line; HSCs, hepatic stellate cells. “anti” indicates anti-fibrotic; “pro” indicates pro-fibrotic. Key lncRNAs and circRNAs: their mechanisms of action in fibrosis of the liver and heart. CFs, cardiac fibroblasts; mCFs, mouse CFs; rCFs, rat CFs; HSC, hepatic stellate cells; mHSC, mouse HSC; LX-2, human hepatic stellate cell line; HSCs, hepatic stellate cells. “anti” indicates anti-fibrotic; “pro” indicates pro-fibrotic. circRNAs CircRNAs represent a distinct class of noncoding RNA molecules abundantly present in mammalian cells [185]. They are formed through a process known as back-splicing, resulting in a covalently closed, single-stranded loop structure [185], [186]. This circular configuration confers resistance to degradation by RNA exonucleases, giving circRNAs remarkable stability in comparison to their linear counterparts [187], [188]. Owing to this stability, they are increasingly explored as candidate circulating biomarkers in preclinical research [189], [190], [191]. circRNAs participate in a wide range of cellular processes, including regulation of transcription and splicing, modulation of protein and RNA interactions, and serving as molecular scaffolds [192]. One of the most characterised functions of circRNAs is their ability to act as miRNA decoys [193], [194]. By binding to miRNAs via specific sites, circRNAs can inhibit miRNA-mediated gene silencing [195]. This interaction prevents degradation of the target mRNAs, thereby increasing their translation into proteins [196]. Growing evidence indicates that circRNAs actively participate in the pathogenesis of fibrosis in multiple organs [197], [198]. Overview of challenges in the detection and interpretation of circulating ncRNAs Circulating ncRNAs have gained attention as candidate biomarkers because they are relatively stable in biofluids and can be readily detected, allowing changes in their abundance to inform disease diagnosis and progression [9], [88], [89], [90]. However, their clinical translation remains limited by several pre-analytical and analytical constraints. First, rigorous sample isolation and purification are essential, as RNA integrity and purity determine detection accuracy [199]. Pre-analytical variables such as hemolysis, serum vs plasma selection, and the type of anticoagulant used can substantially alter circulating ncRNA profiles, thereby confounding interpretation. In particular, hemolysis can artificially increase erythrocyte-enriched ncRNAs such as miR-16-5p, making commonly used endogenous controls unreliable in blood-derived biofluids [200]. ncRNA content also differs between plasma and serum [201], [202]. McDonald et al. demonstrated that plasma samples generally contained higher levels of circulating miRNAs than serum, largely due to miRNA release from platelets and other blood cells during the coagulation process in serum preparation [201]. Anticoagulants also require careful standardisation: EDTA or citrate are generally preferred for RT-qPCR-based circulating ncRNA analyses, whereas heparin can inhibit PCR amplification, potentially leading to inaccurate ncRNA quantification [203]. Therefore, consistent reporting of blood collection tubes, anticoagulant type, processing time, and storage conditions is essential to improve reproducibility and comparability across studies (Fig. 3). Technical and translational challenges limiting the clinical implementation of circulating ncRNA biomarkers. Second, data normalisation is challenging due to the lack of a universally stable endogenous reference in blood-derived biofluids. Commonly used miRNA controls (miR-16-5p and U6) can be affected by disease-related instability, while alternative approaches, including spike-ins or volume-based, have not yet been uniformly standardised [204], [205]. Interpretation of circRNA studies is also limited by normalisation issues. circRNA quantification is highly dependent on primer design, back-splice junction detection, RNA quality, and analytical pipelines [206], [207]. In biofluids, commonly used controls such as U6, GAPDH, ACTB, or 18S rRNA may be unsuitable because they are intracellular and/or linear RNAs and may not reflect circRNA stability, extraction efficiency, or RNase resistance. Similarly, external spike-ins can monitor extraction variability but do not fully mimic endogenous circular structure or back-splice junction amplification. Therefore, reported disease-associated changes in circulating circRNAs, including fibrosis-related circRNAs, may partly reflect technical normalisation bias rather than true biological variation. Finally, interpreting circulating ncRNAs is complicated by their mixed origins, as they may be released from multiple tissues and cell types. Overall, without standardised pre-analytical protocols, validated normalisation strategies, and tissue-of-origin clarification, circulating ncRNAs remain promising but not yet fully reliable biomarkers (Fig. 3). ncRNA-mediated mechanisms in Heart and Liver fibrosis Fibrosis in the heart and liver is increasingly viewed as part of an interconnected, multisystem remodelling process rather than isolated organ pathology [1], [4], [5], [208]. miRNAs, lncRNAs and circRNAs act as fine-tuners of gene expression and have emerged as critical regulators of molecular pathways involved in fibrosis, potentially linking pathological remodelling mechanisms between the two organs [88], [197], [209]. Numerous miRNAs, lncRNAs, and, more recently, circRNAs have been implicated in the regulation of key fibrosis-associated signalling pathways, such as TGF-β, WNT, TNF-α, NF-κB, and YAP/TAZ, across multiple organ systems, including liver and heart [210]. Understanding these unifying, but also unique ncRNA-mediated mechanisms may pave the way for new diagnostic and therapeutic strategies aimed at managing multisystem fibrotic disease [1], [4], [6]. In this section, we summarise overlapping ncRNA implicated in both hepatic and cardiac fibrosis, highlighting where regulation is convergent, divergent or mixed and mapping these findings to the underlying pathways. miRNAs and their mechanisms in heart and liver fibrosis miR-21 miR-21 is among the most consistently reported pro-fibrotic miRNAs and is frequently upregulated in both hepatic and cardiac remodelling, making it a plausible shared regulator across organs [211], [212], [213]. In both liver and heart, miR-21 is increased during the activation of HSCs and CFs and converges on common pro-fibrotic signalling programmes. These include TGF-associated pathways, ERK/MAPK-linked fibroblast activation [100], [102], [104], [107], and repression of PTEN [100], [214] (Fig. 4). However, despite these commonalities, the precise molecular mechanisms and cellular contexts still differ between tissues, requiring independent validation prior to broader generalisation (Table 3). Role of miR-21 in liver and heart fibrosis. In both the liver and heart, miR-21 promotes hepatic stellate cell (HSC) and cardiac fibroblast (CFs) activation through common pro-fibrotic pathways, including ERK1 signalling, the TGF-β pathway [100], [102], [104], [107] and repression of PTEN [100], [214]. The red colour of the miR-21 box represents its reported pro-fibrotic role in both liver and heart fibrosis. Yellow colour highlights potentially shared regulatory pathways. Dashed lines indicate that the depicted pathways are context- and model-dependent, reflecting the fact that the regulatory relationships shown have been reported across heterogeneous experimental systems and may not operate uniformly across all fibrotic contexts. Circulating miR-21 tends to rise with fibrosis burden in both CVD and CLD cohorts [85], [215]. Elevated miR-21 has been reported in HF [216], [217], [218], and miR-21-5p was associated with atrial fibrosis severity in atrial fibrillation (AF) patients (n = 175) [215]. In liver disease, serum miR-21 increases with fibrosis severity in chronic hepatitis B (CHB, n = 28) [85], and is higher in more advanced fibrosis in NAFLD, correlating with established fibrosis indices [97]. Overall, these studies position miR-21 as a common pro-fibrotic effector in both hepatic and cardiac settings, with convergent upregulation in tissue and circulation (Fig. 4, Fig. 7). However, miR-21 broad elevation in other pathological conditions, including cancer, limits its specificity for fibrosis [219]. Convergent and divergent roles of ncRNAs in liver vs heart fibrosis. The upper part of the schematic summarises ncRNAs with convergent effects (the same direction - either pro-fibrotic or anti-fibrotic - in both liver and heart). The lower part of the schematic summarises divergent effects (opposite directions, i.e., pro-fibrotic in one organ but anti-fibrotic in the other). The red colour of ncRNAs boxes represents reported pro-fibrotic role, green - anti-fibrotic. Yellow colour highlights potentially shared regulatory pathways. Dashed lines indicate that the depicted pathways are context- and model-dependent, reflecting the fact that the regulatory relationships shown have been reported across heterogeneous experimental systems and may not operate uniformly across all fibrotic contexts. miR-22 miR-22 is a conserved miRNA implicated in both cardiac and hepatic remodelling but showing context-dependent and sometimes opposing effects [220], [221]. One possible explanation is that miR-22 gives rise to two mature strands, miR-22-3p and miR-22-5p [92], yet studies do not always report which strand was measured, contributing to inconsistent findings. In the heart, mechanistic work implicates miR-22-3p in cell-intrinsic effects. In MI-model, miR-22 has been reported to promote fibrosis via regulation of caveolin-3 [111]. Notably, caveolin-3 is a muscle-specific protein, suggesting that this mechanism is likely heart-specific and absent in the liver [111]. Surprisingly, miR-22 was also described as anti-fibrotic, suppressing TGFβRI expression [114] and inhibiting PTEN/Akt/mTOR pathway [115], while miR-22-3p also targeted PTAFR [112] in post-MI remodelling. These divergent findings likely stem from differences in models and disease stage, as well as from the dominant downstream target network engaged in a given context. In the liver, miR-22 has also emerged as a critical regulator of inflammation and fibrosis [220]. Most studies describe miR-22 as anti-fibrotic in hepatic tissue. miR-22-3p and miR-29a-3p act synergistically to inhibit HSC activation [109], and miR-22-3p attenuates liver fibrosis by regulating cytohesin-3 [108]. Available findings are compatible with miR-22 acting in both organs as a modulator of fibrotic signalling (particularly PI3K/Akt-associated pathways) yet exhibiting mixed pro- or anti-fibrotic effects (Table 3) - within the same organ across studies (Fig. 5) [109], [115]. Mixed pro- and anti-fibrotic roles of miRNAs in liver vs heart. Schematic summarising miRNAs reported to exert opposing (pro- or anti-fibrotic) effects in either hepatic or cardiac tissue. The red colour of miRNAs boxes represents reported pro-fibrotic role, green – anti- fibrotic. miR-22, miR-29, miR-122 and miR-214 are depicted with representative pathways or downstream targets through which they have been linked to either attenuation or promotion of fibrotic remodelling in each organ. Yellow colour highlights potentially shared regulatory pathways. Dashed lines indicate that the depicted pathways are context- and model-dependent, reflecting the fact that the regulatory relationships shown have been reported across heterogeneous experimental systems and may not operate uniformly across all fibrotic contexts. Small cohort studies have reported serum miR-22 levels to be elevated in patients with chronic HF [86]. In addition, plasma miR-22-5p has been explored as an early candidate biomarker in acute MI, strengthening its CVD relevance [83]. Similarly, in patients with NAFLD, serum miR-22-5p was increased [222]. While functional diversity shows that miR-22 regulates multiple fibrotic pathways, it limits its usefulness as a stand-alone biomarker and highlights the need for careful interpretation in both experimental and clinical settings. miR-29 family The miR-29 family (miR-29a/b/c) is widely regarded as a core anti-fibrotic miRNA group (“master-fibromiRNA”), as reduced miR-29 is repeatedly associated with increased ECM gene expression and fibrosis across organs [223]. In liver fibrosis, miR-29 family members are consistently downregulated [117], [224]. Restoring miR-29 attenuates fibrosis both in vitro and in vivo. These effects are partly mediated by a reciprocal loop between miR-29b and Smad3, as well as by targeting PIK3R1 and AKT3 [117], [224]. In the heart, miR-29 is also frequently reported to be anti-fibrotic by targeting factors such as TGF-β, c-FOS, VEGF-A, SH2B3, and DNMT3A [225]. However, in contrast to the liver, some studies also suggest its pro-fibrotic effects through suppression of various components in the Wnt pathway [225] (Fig. 5). Despite these differences, a plausible shared anti-fibrotic mechanism may lie in convergence on cell-cycle control, through suppression of cyclin-CDK activity, CDK6 - in the liver [116], and CDK2 - in the heart [119]. Nonetheless, caution is warranted when extrapolating these findings across organs, as the underlying molecular mechanisms and cellular contexts vary considerably between the tissues. Interpreting or exploiting miR-29 therapeutically requires careful considerations due to its mixed pro-fibrotic anti-fibrotic effects in the heart. Fig. 6 and Table 3 summarise the major molecular pathways of miR-29 underpinning liver and cardiac fibrosis. Role of miR-29 family in liver and heart fibrosis. The miR-29 family exerts broad anti-fibrotic effects (which are represented by the green colour of the miRNA box) in both the liver and heart by regulating multiple signalling pathways involved in inflammation, cell cycle control, and fibroblast activation [116], [117], [119]. However, in the heart, it could also promote CMs hypertrophy, contributing to cardiac remodelling [118]. Dashed lines indicate that the depicted pathways are context- and model-dependent, reflecting the fact that the regulatory relationships shown have been reported across heterogeneous experimental systems and may not operate uniformly across all fibrotic contexts. Preclinical observations broadly support an association between reduced circulating miR-29 and fibrotic remodelling in both organs [224], [226], [227], [228]. Importantly, miR-29b-3p and miR-29c-3p were significantly upregulated in univentricular heart patients and were strongly correlated with advanced liver fibrosis/cirrhosis scores. Expression of both molecules were higher in group without Fontan palliation, reflecting worse prognosis [229]. Circulating miR-29a was decreased in advanced liver cirrhosis [224], metabolic dysfunction-associated steatotic liver disease (MASLD) [227]. Conversely, circulating miR-29a was upregulated in HCM patients and correlated with left ventricular (LV) hypertrophy and fibrosis [84]. miR-29b was reported reduced in HF and AF, with further decrease when both conditions coexist [228]. To conclude, the miR-29 family is an emerging candidate biomarker for both cardiac and liver fibrosis, based on preclinical observations and preliminary clinical associations. However, current evidence is limited by small cohort sizes, a lack of validation cohorts, and often poor adjustment for confounding variables (e.g., age, sex, comorbidities, medications). Additional limitations include a lack of organ specificity (given that circulating levels may derive from multiple tissue sources, not only the liver and heart) [223], as well as stage- and aetiology-dependent or non-linear behaviour across disease progression [226]. Consequently, miR-29 is likely to perform better within multi-marker panels rather than as a standalone test [230]. miR-34a miR-34a is a pro-fibrotic regulator implicated in pathology through regulation of cellular proliferation, differentiation, and apoptosis [231]. Studies support its relevance in cardiac remodelling, including miR-34a-mediated repression of SKI, an inhibitor of TGF-β-signalling [121]. Moreover, plasma miR-34a levels were higher in patients who develop LV remodelling after AMI (n = 116) compared with non-remodelling patients (n = 243), and elevated miR-34a is associated with a risk of HF [232]. Complementary data show miR-34a is elevated in liver fibrotic tissue, and its suppression reduces HSC activation, proliferation and migration [233]. miR-34a implies pro-fibrotic signalling via SIRT1/p53 positive feedback loop [120]. Circulating miR-34a is elevated in both MASLD (n = 50) and MASLD plus T2D patients (n = 48) compared to the control group (n = 49) [234]. Serum levels of miR-34a (n = 28) were also higher in NAFLD patients than in healthy controls (n = 36) [99]. Collectively, these studies support miR-34a as a cross-organ pro-fibrotic effector with convergent pathogenic roles in cardiac and hepatic fibrosis (Fig. 7). However, as with any proposed cross-organ mechanism, these convergent roles require further independent mechanistic validation to fully understand context-specific contributions (Table 3). Biomarker claims should be also interpreted with caution, as available studies are based on modest cohort sizes and heterogeneous designs, underscoring the need for larger, longitudinal, and standardised assays before clinical translation [99], [232], [234]. miR-122 miR-122 is one of the most abundant hepatic miRNAs, accounting for ∼70% of total liver miRNAs [124]. It plays a central role in lipid and cholesterol metabolism, hepatocyte differentiation, and liver homeostasis [235], [236], [237]. Most studies support the anti-fibrotic role of miR-122 in the liver [235], [236], [237]. Its restoration suppresses HSC proliferation and ECM production by targeting P4HA1, a key collagen maturation enzyme [122] (Fig. 5). Moreover, loss of miR-122a in mice leads to fibrosis partially through derepression of transcription factor KLF6, which participates in the activation of the TGF-β pathway [124]. Evidence for the involvement of miR-122 in cardiac fibrosis is context-dependent [238]. In myocardial biopsies from patients with severe aortic stenosis, miR-122 expression is reduced [127]. Overexpression of miR-122-5p in TGF-β1-treated CFs inhibited Smad pathway [126]. This converges with liver findings, where loss of miR-122a leads to TGF-β pathway activation [124] (Fig. 5). Conversely, in hypertension models, miR-122-5p exerted a pro-fibrotic role by disrupting the elabela/apelin-ACE2-GDF15 axis [125]. However, the study reports downregulation of GDF15, which is contradictory to reports that hypertension can induce GDF15 expression as a protective response and might be explained in part by different experimental models [125]. Overall, miR-122 is largely anti-fibrotic in the liver, while its cardiac effects are context-dependent and require further study. For this reason, miR-122 is better categorised as ncRNAs with mixed roles in fibrosis (Fig. 5, Table 3). Circulating miR-122 has been investigated in relation to multiple liver diseases [239]. A longitudinal study of NAFLD showed that miR-122 dynamically predicts outcomes in patients with advanced fibrosis [240]. A meta-analysis further confirmed that serum miR-122 is upregulated across all stages of NAFLD severity [241]. Serum miR-122 levels were also increased in biopsy-confirmed MASLD, showing better diagnostic performance than alanine aminotransferase (ALT) [242]. In CVD, higher levels of circulating miR-122-5p following transcatheter aortic valve replacement predict worse recovery of ejection fraction [243]. Although promising, these findings would benefit from confirmation in additional, larger cohorts and with clearer links between circulating levels and remodelling mechanisms. miR-133a miR-133a is a muscle-enriched miRNA that nonetheless appears repeatedly across cardiac–hepatic injury settings, consistent with a role as a cross-organ modulator of remodelling and a potential readout of tissue stress [128], [132]. In the heart, miR-133a demonstrates anti-fibrotic effects in diabetic cardiomyopathy (DCM), where its overexpression prevents ERK1/2 and SMAD-2 phosphorylation and suppresses collagen [131], [244]. Similar anti-fibrotic actions are observed following MI, where miRNA-133a suppresses TGF-β1 [132], and during ischaemia-perfusion (I/R) injury, where miR-133a-3p limits fibrosis via SOX5-dependent pathways [130]. In the liver, miR-133a displays parallel anti-fibrotic functions by regulating TGF-β-pathway and collagen production [128], [129] (Fig. 7). Across CVD cohorts, circulating miR-133a has repeatedly been reported to increase with myocardial injury [245], [246], [247]. It correlates with HF severity markers (LV end-systolic volume and NT-proBNP) in patients undergoing cardiac resynchronisation therapy [248], and is elevated in both ST- and non-ST-elevation MI [245], [246], [247]. It has also been shown to reflect myocardial steatosis in type 2 diabetes (T2D) [98]. In CLD (n = 70) serum miR-133a was elevated compared with healthy controls (n = 19), with the higher increase observed in individuals with earlier-stage fibrosis [128]. Collectively, miR-133a represents a candidate indicator of cardiac and hepatic injury with convergent anti-fibrotic roles in both cardiac and hepatic disease (Fig. 7) [129], [132]. However, further studies are essential to validate it. miR-210 miR-210 is the master microRNA of hypoxia response downstream of HIF-1/2, linking low-oxygen stress to inflammatory and fibrotic remodelling [249]. Its effects depend on the targeted cell type and whether hypoxia signalling drives maladaptive activation or protective stress adaptation [249], [250], [251]. In the heart, evidence largely support a pro-fibrotic role of miR-210, particularly in pressure overload [251], [252]. A cross-sectional study of hypertensive HF patients reported associations of miR-210-3p expression with NT-proBNP, sST2, and Gal-3 [253]. Plasma profiling studies also show miR-210-3p correlation with HF severity [254]. However, these data should be interpreted as associative and hypothesis-generating until replicated across larger, longitudinal cohorts. Indeed, hypoxia-responsive miRNAs can be influenced by many confounders (injury burden, comorbid ischaemia, renal function, and sampling time) that limit specificity for fibrosis [249], [253], [254]. In the liver, miR-210 has been identified within HIF-dependent miRNA networks governing stress responses in HSC [255]. In metabolic dysfunction–associated steatohepatitis (MASH), miR-210-3p can be anti-fibrotic by promoting ferroptosis of activated HSC, and its upregulation in vivo markedly attenuated liver fibrosis [256]. Thus, a reasonable hypothesis is that miR-210 is a miRNA with mixed effects: it is pro-fibrotic when it sustains inflammation and CFs activation in hypoxic cardiac stress but can be anti-fibrotic in the liver when it shifts the balance by reducing the activated HSC pool within hypoxia-regulated injury programmes [252], [255], [256]. miR-214 miR-214 is a multifunctional miRNA implicated in liver and heart fibrosis. In the liver, miR-214 exhibits both anti- and pro-fibrotic actions, reflecting strong context-dependent effects [257]. On the anti-fibrotic side, miR-214 is reduced during HSC activation, and its overexpression attenuates fibrogenesis via repression of MECP2 [134]. However, whole-liver miR-214 overexpression also promotes steatosis and inflammation, suggesting cell type-specific effects [134]. Limitations include the unclear mechanisms driving its early downregulation and the unresolved pathways linking miR-214 to alterations in hepatic lipid metabolism. miR-214 expression could be driven by the transcription factor Twist1, yet downstream effects diverge [135], [136]. One study showed that Twist1-induced miR-214 suppressed CCN2 and exerted an anti-fibrotic effect [136], whereas another demonstrated that the Twist1-miR-214-3p axis promoted HSC activation by targeting Sufu and enhancing Hedgehog signalling [135]. Methodological differences ( [136] normalised miR-214 to GAPDH mRNA, [135] - to U6 RNA) may contribute to discrepancies, but clinical data link severe fibrosis to increased miR-214 and reduced Sufu, supporting a pro-fibrotic Hedgehog-related mechanism [258]. Circulating miR-214 has been discussed as a possibly emerging biomarker in CLD [259], [260], with higher levels in whole blood reported in patients with liver fibrosis [87]. Elevated serum miR-214 was observed in patients with biliary atresia with severe fibrosis, compared to those with mild fibrosis [258]. However, these findings should be interpreted cautiously, as mechanistic studies show compartment-specific, and sometimes opposing regulation, and the reported cohorts remain modest in size. In CVD, miR-214 data are likewise inconsistent and influenced by treatment status (e.g., higher circulating levels in HF patients on β-blockers, with no clear separation from healthy controls in small cohorts) [261] (Fig. 5, Table 3). Mechanistically, miR-214 can be anti-fibrotic (e.g., via NLRC5 [137], EZH1/2 [138]), but also pro-fibrotic through c-Ski repression [140] or activation of ERK1/2 signalling [143]. Overall, miR-214 is best regarded as a context-dependent regulator with mixed effects in fibrosis (Fig. 5). Future studies should explicitly resolve cell origin, strand specificity, and pathway engagement using standardised gain- and loss-of-function approaches with validated targets and matched human data to improve interpretability and relevance [259]. lncRNAs and their mechanisms in heart and liver fibrosis H19 The lncRNA H19 represents a clear example of a pro-fibrotic regulatory RNA implicated in both hepatic and cardiac fibrogenesis, and, notably, detectable in the circulation [262], [263], [264]. Serum H19 was upregulated in patients with MAFLD and was positively associated with liver stiffness [262]. Likewise, circulating H19 distinguished patients with pulmonary hypertension, serving as a predictor of long-term survival [263]. Mechanistically, H19 promotes liver fibrosis through several pathways, including activation of TGF-β1 signalling (H19/miR-148a axis) [160], EZH2-mediated chromatin remodelling [161], and ERK-dependent signalling regulated by DNMT1-H19 interaction [162]. Additionally, a Sox9/H19 axis has been identified whereby Sox9-driven H19 upregulation promotes hepatocyte apoptosis and subsequent HSC activation [265]. Cardiac studies show notable parallels [264]. H19 promotes CFs proliferation through DUSP5/ERK1/2 modulation [164]. It also promotes fibrosis by suppressing miR-29a-3p/miR-29b-3p-VEGFA/TGF-β axis [163] and by sponging miR-455 to increase CTGF expression [165]. Interestingly, H19-dependent fibrosis mechanisms also include let-7 family regulation across organs [264]. Together, these findings indicate that H19 integrates metabolic, epigenetic and signalling pathways across tissues and is a shared mechanistic driver and candidate biomarker in both liver and cardiac fibrosis. H19 is pro-fibrotic in both liver and heart, converging on the TGF-β and ERK1/2 pathways (Fig. 7). However, the reported downstream targets and interacting partners differ between cellular compartments. Observations of circulating H19 (and other circulating lncRNAs) should also be interpreted with caution: current studies are limited by small cohorts and heterogeneity in sample collection and processing. Larger, standardised, multi-centre validation are required before clinical translation. MALAT1 MALAT1 (metastasis-associated lung adenocarcinoma transcript 1) is a pro-fibrotic lncRNA involved in both heart and liver fibrosis [159], [177]. In the heart, it promotes post-MI remodelling by regulating TGF-β1 via miR-145 [176] and mediates DCM via TGF-β1/Smad and miR-141/NLRP3 pathways [177]. In CVD patients, MALAT1 is recognised as a possible biomarker of AMI, although its specificity, reproducibility and long-term association with cardiac fibrosis remains to be better characterised [157], [266]. In liver fibrosis, MALAT1 drives HSC activation and inflammation through the miR-181a/TLR4/NF-κB axis [173]. MALAT1 suppression in HSC reduced myofibroblast markers and restored SIRT1, consistent with patient data showing lower SIRT1 and higher MALAT1 levels in cirrhosis [267], [268]. Other mechanisms in liver fibrosis progression involved MALAT1 regulating CXCL5 and Rac1 expression, influencing the proliferation, cell cycle, and HSC activation [174], [175]. Serum MALAT1 levels were reported to be higher in fibrotic patients (n = 151) vs. healthy controls (n = 93) and were positively correlated with fibrosis severity [173]. Similarly, in cirrhosis, circulating MALAT1 and miRNA-181a levels might serve as candidate biomarkers [267]. Overall, MALAT1 emerges as a convergent multifaceted driver of pro-fibrotic signalling across organs (Fig. 7). MEG3 The lncRNA MEG3 (maternally expressed gene 3) is generally downregulated in fibrotic liver tissues and activated HSC. Restoration of MEG3 suppresses HSC activation and inhibits fibrogenesis [166]. In metabolic liver disease, MEG3 exerts protective effects upregulating SIRT6 (via EXH2 ubiquitination) [170]. These findings support a model in which MEG3 primarily functions as an anti-fibrotic regulator in the liver. Serum MEG3 was reduced in patients with hepatic fibrosis and decreased with disease severity, distinguishing mild from advanced fibrosis [169]. Similarly, MEG3 in serum was lower in individuals with chronic hepatitis B (CHB, n = 100, vs. 80 healthy controls) and showed a negative correlation with liver fibrosis severity [269]. In contrast, in the heart MEG3 promotes pathological remodelling. Its in vivo inhibition after transverse aortic constriction improved diastolic function [153]. Similarly, MEG3 knockdown protected against HF in mice by reducing oxidative stress, apoptosis and fibrosis through the miRNA-129-5p/ATG14/Akt axis [171]. In DCM, MEG3 was also upregulated and worsened inflammation, fibrosis and hypertrophy via miR-223/NLRP3, while MEG3 silencing reduced these changes [172]. Elevated serum MEG3 levels are associated with reduced heart rate variability (HRV) in HF patients (n = 45 – HF with ventricular arrhythmia (VA) vs. 85 HF without VA). Combined MEG3-HRV assessment improved prediction of VA [270]. Differences in MEG3 function across organs likely reflect (i) cell-type specificity in available miRNAs, RNA-binding proteins and transcriptional factors that alter MEG3’s effective targets [166], [171]; (ii) epigenetic landscape - MEG3 is transcribed from the imprinted DLK1–DIO3 locus, where DNA methylation controls its expression [271]; (iii) subcellular localisation - nuclear vs. cytoplasmic localisation affects whether MEG3 primarily regulates transcription/chromatin or acts as a miRNA sponge. However, reports on the predominant localisation of MEG3 are diverse even within the same cell type. MEG3 has been detected in both compartments in cardiomyocytes [272], [273], and explicit predominant localisation is not clearly established. The opposing effects reported for MEG3 modulation may be attributable to differences in MEG3 isoform expression. Most studies did not distinguish among MEG3 isoforms; an exception is Yu et al. [166], which addressed isoforms explicitly. During liver and heart fibrosis, distinct MEG3 isoforms may be expressed, and MEG3 isoform usage is known to be tissue-specific. For example, human fetal liver expresses all 12 MEG3 isoforms, which share common exons 1–3 and 8–10 but differ through alternative splicing of exons 4–7 [274]. However, limited knowledge of heart-specific MEG3 isoforms constrains direct comparisons of MEG3 modulation between cardiac and hepatic tissues. Practically, mechanistic findings from one organ or cell type should not be extrapolated without direct validation. Overall, MEG3 exerts mainly protective roles in the liver - suppressing HSC activation, reducing collagen production, and serving as a candidate biomarker of fibrosis severity in small cohort preclinical studies. Conversely, in the heart, MEG3 generally promotes pathological remodelling, contributing to fibrosis, inflammation, hypertrophy, apoptosis, and impaired cardiac function, with its inhibition consistently improving cardiac outcomes (Fig. 7). Overall, MEG3 has divergent roles in fibrosis between the heart and the liver. While the cited studies robustly support these organ-specific trends, it is crucial to recognise that differences in experimental models, the expression of distinct MEG3 isoforms, or varying cellular contexts within each organ may contribute to these divergent findings. As highlighted by the complexity of MEG3 isoform expression, mechanistic findings from one organ or cell type should not be extrapolated without direct validation. Therefore, when generalising the observed pro- or anti-fibrotic roles of MEG3, it is important to consider these underlying complexities rather than assuming a universal mechanism across tissues. NEAT1 NEAT1 (Nuclear Enriched Abundant Transcript 1) is described as pro-fibrotic in both the liver and the heart [275]. NEAT1 promotes the activation of HSC by sponging miRNA-122 to enhance KLF6 expression [178], miR-139-5p to activate the β-catenin/SOX9/TGF-β1 pathway [179], miR-148a-3p and miR-22-3p to increase Cyth3 [108]. Additional mechanisms include modulation of inflammation in NAFLD through the miR-506-GLI3 axis [276], promotion of alcoholic steatohepatitis-related fibrosis via miR-129-5p/SOCS2 [277], and deterioration of NASH-related fibrosis through miR-129-5p/PEG3 [180]. Thus, NEAT1-targeted inhibition may represent a potential therapeutic strategy across different liver fibrosis aetiologies [108], [180]. As a blood-based molecule, at least two studies demonstrated that lncRNA NEAT1 is expressed in patients with NAFLD in serum and peripheral blood monocytes (PBMCs) [278], [279], [280]. NAFLD patients with increased NEAT1 in PBMCs also showed a positive correlation with hepcidin, a liver-produced hormone [280]. In the setting of cardiac fibrosis, NEAT1 plays a role through alternative pathways [275]. In AMI, NEAT1 expression was significantly elevated in PBMCs of affected patients (n = 25) compared with healthy individuals (n = 25) and those with stable angina (n = 25) [281]. Higher NEAT1 levels were associated with lower circulating markers of myocardial injury and inflammation (CK-MB, cTnI, TNF-α, IL-6, IL-1β) and a higher LV ejection fraction [281]. Overall, NEAT1 is a ncRNA with convergent effects in heart and liver fibrosis (Fig. 7, Table 3). The literature on circulating lncRNAs is expanding rapidly, but evidence remains limited by small cohorts in pre-clinical studies, variable pre-analytical handling (plasma vs. serum, haemolysis), differences in RNA extraction and normalisation methods, and incomplete separation of vesicle-associated vs. protein-bound fractions. These factors can generate study-to-study variability and overinterpretation risk if not controlled, highlighting the need for rigorous methodological standardisation and large-scale validation before clinical translation. circRNAs and their mechanisms in heart and liver fibrosis circHIPK3 Only a limited number of circRNAs have been validated as fibrosis-related in both heart and liver. Among them circHIPK3 (circular RNA homeodomain interacting protein kinase 3; isoforms hsa_circ_0000284 and mmu_circ_0001052) is a notable example [9], [198]. In the heart, it promotes CFs proliferation, migration and phenotype switching in response to hypoxia, by sponging anti-fibrotic miRNAs (e.g., miR-152-3p) [184], [282]. Inhibition of circHIPK3 consistently attenuates cardiac fibrosis in both in vitro and in vivo models, underscoring its relevance as a potential therapeutic target [282]. Several studies indicate that circHIPK3 might have biomarker potential in CVD. Exosomal circHIPK3 differentiated chronic coronary syndrome patients from non-cardiac chest pain [283]. The circulating circHIPK3/miR-221 ratio identifies recent cerebrovascular events [284], and circHIPK3 levels have also been linked to hypertension risk [285]. In the liver, circHIPK3 expression in serum was upregulated in patients with liver fibrosis and liver cirrhosis (n = 120) vs. healthy controls (n = 120) and was positively correlated with serum ALT and aspartate aminotransferase (AST) levels [189]. circHIPK3 is upregulated in fibrotic liver tissue and positively correlates with TGF-β1 and COL1A1 [189]. Overall, circHIPK3 is a pro-fibrotic circRNA with convergent effects (Fig. 7). It is also upregulated in both heart and liver disease, and it might hold potential as a circulating biomarker and therapeutic target across cardiovascular, hepatic, and metabolic disorders [184], [189], [282]. Table 5 provides an overview of ncRNAs proposed as potential biomarkers in liver- and cardiac diseases associated with fibrosis. Table 5Circulating ncRNAs as emerging candidate biomarkers of cardiac and liver disease progression.Primary affected organBiofluidModulationFibrosis ROC/correlationExperimental groupControl groupRef.ConditionnConditionnmiR-21Liverserumup+CHB, fibrosis stage F47CHB, fibrosis stage F1-F321[85]serumup+NAFLD, fibrosis stage F3-F429NAFLD, fibrosis stage F0-F321[97]HeartserumupN/AHF80HC40[218]serumup+AF with LVA fibrosis110AF with no/minor LVA fibrosis65[215]plasmaupN/AHF903non-HF807[216]plasmaupN/AHF9HC8[217]miR-22HeartserumupN/AChronic HF30HC30[86]plasmaupN/AAMI35Non-AMI55[83]LiverserumupN/ANAFLD44Non-NAFLD10[222]miR-29 familyLiverserumdown+Mild/severe CLD67HC17[224]serumdownN/ACHB with fibrosis Cirrhosis91HC12[226]serumdown+MASLD30HC20[227]HeartplasmadownN/AAF33Non-CVD controls30[228]plasmaup+HCM41HC41[84]miR-34aLiverplasmaupN/AMASLD50Non-MASLD controls49[234]serumupN/ANAFLD28HC36[99]HeartplasmaupN/AAMI, LV remodelling116No LV remodelling243[232]miR-122LiverserumdownN/ANAFLD and liver cancer7NAFLD54[240]HeartplasmaupN/APost-TAVI, EF < 50%15Post-TAVI, EF> 50%15[243]miR-133aLiverserumupN/AChronic hepatic disease70HC19[128]HeartplasmaupN/ASTEMI25HC11[247]miR-210LiverplasmaupN/ACHB, early responders37CHB, non-early responders38[286]serumup-CHB115HC20[287]HeartplasmaupN/ACHF30HC36[254]serumupN/AAF50SR50[288]serumupN/AHFrEF100HFpEF170[253]miR-214Liverwhole bloodup+NAFLD with liver fibrosis19HC7[87]serumup+BA with severe fibrosis17BA with mild fibrosis11[258]HeartserumupN/AHF receiving β-blockers35HF no β-blockers20[261]H19Liverserumup+MAFLD81Non-MAFLD40[262]HeartplasmaupN/APulmonary hypertension:existing cohort/validation cohort52/75Non-hypertensive controls:existing cohort/validation cohort57/54[263]MEG3Liverserumdown+CHB100HC80[269]HeartserumupN/AHF with VA45HF without VA85[270]MALAT1Liverserumup+Liver fibrosis151HC93[173]HeartPBMCupN/AAMI132HC104[157]PBMCupN/AAMI160Angina pectoris50[266]NEAT1LiverserumupN/ANAFLD10HC10[279]PBMCsupN/ANAFLD119HC100[280]HeartPBMCsupN/AAMI25HC Stable angina2525[281]circHIPK3Liverserumup+Liver cirrhosis120HC120[189]HeartplasmaupN/AChronic coronary syndrome135Non-cardiac chest pain83[283]Fibrosis ROC/Correlation: +, the study reported a statistically significant correlation (Pearson/Spearman) with a direct fibrosis marker (e.g., Ishak score) and/or a significant ROC curve analysis (AUC > 0.7) for fibrosis detection or staging; -, the study reported non-significant correlation or ROC results; N/A, no correlation with a direct fibrosis marker or ROC curve analysis was performed in the study.AF: Atrial fibrillation; AMI: Acute myocardial infarction; BA: Biliary atresia; CHB: Chronic hepatitis B; CHM: Congestive hepatopathy (cardiac hepatopathy); HC: Healthy control; HF: Heart failure; HFpEF: Heart failure with preserved ejection fraction; HFrEF: Heart failure with reduced ejection fraction; LVA: low voltage areas; LVEF: Left ventricular ejection fraction; MASLD: Metabolic dysfunction-associated steatotic liver disease; NAFLD: Nonalcoholic fatty liver disease; SR: Sinus rhythm; STEMI: ST-segment elevation myocardial infarction; TAVI: Transcatheter aortic valve implantation. Circulating ncRNAs as emerging candidate biomarkers of cardiac and liver disease progression. Fibrosis ROC/Correlation: +, the study reported a statistically significant correlation (Pearson/Spearman) with a direct fibrosis marker (e.g., Ishak score) and/or a significant ROC curve analysis (AUC > 0.7) for fibrosis detection or staging; -, the study reported non-significant correlation or ROC results; N/A, no correlation with a direct fibrosis marker or ROC curve analysis was performed in the study. AF: Atrial fibrillation; AMI: Acute myocardial infarction; BA: Biliary atresia; CHB: Chronic hepatitis B; CHM: Congestive hepatopathy (cardiac hepatopathy); HC: Healthy control; HF: Heart failure; HFpEF: Heart failure with preserved ejection fraction; HFrEF: Heart failure with reduced ejection fraction; LVA: low voltage areas; LVEF: Left ventricular ejection fraction; MASLD: Metabolic dysfunction-associated steatotic liver disease; NAFLD: Nonalcoholic fatty liver disease; SR: Sinus rhythm; STEMI: ST-segment elevation myocardial infarction; TAVI: Transcatheter aortic valve implantation. Exosome-associated ncRNAs as candidate mediators of cardio-hepatic crosstalk Extracellular vesicles (EVs) is an umbrella terms, which unites all types of lipid bilayer vesicles released from cells. A specific form of such vesicles, exosomes, ranging from approximately 30–150 nm in diameter, is frequently studied as containing and transporting miRNA cargo [289]. They originate from the endosomal pathway through inward budding of the endosomal membrane to form intraluminal vesicles (ILVs) within multivesicular bodies (MVBs) [290]. Exosome formation depends largely on the ESCRT (endosomal sorting complex required for transport) machinery, which sorts cargo, promotes membrane budding and scission to form ILVs within MVBs, followed by Rab- and SNARE-mediated trafficking and fusion of MVBs bodies with the plasma membrane to release exosomes [291]. Exosomes facilitate intercellular communication by transporting bioactive cargo, including proteins, lipids, and ncRNAs [290]. The selective sorting of ncRNAs into exosomes is regulated by lipid- and protein-mediated mechanisms, including the neutral sphingomyelinase 2 (nSMase2)-dependent pathway and RNA-binding proteins (RBPs) such as hnRNPA2B1, which recognise specific sequence motifs (EXOmotifs) in miRNAs and guide their selective loading into exosomes [292], [293]. This section outlines a hypothesis-driven and currently non-definitive perspective on exosome-associated ncRNAs in cardio-hepatic crosstalk. At present, direct evidence demonstrating causal exosome-mediated ncRNA transfer between the liver and heart in humans is limited. Therefore, the studies discussed below should be treated mainly as supportive of biological plausibility rather than as proof of established inter-organ transfer. Importantly, shared detection or upregulation of the same ncRNA in the heart, liver, or circulation should not be interpreted as evidence of endocrine or paracrine ncRNA transfer. Demonstrating true inter-organ communication would require substantially stronger evidence, including tissue-of-origin tracing, labelled vesicle tracking, functional transfer experiments showing recipient-organ effects, and paired longitudinal measurements of donor tissue, circulation, and target tissue within the same experimental or clinical setting [294]. Another crucial challenge lies in demonstrating whether the physiological concentrations of specific ncRNAs and the vesicles themselves are sufficiently high in vivo to meaningfully contribute to cardiac remodelling processes [295]. Under stress, hepatocytes, HSC, CMs, and CFs, release EVs enriched in ncRNAs. EVs may be taken up by distant cells and could potentially modulate inflammation, metabolism, and fibrosis, providing a plausible explanation why distinct ncRNA signatures correlate with fibrosis severity in both liver and cardiac disease [296], [297]. However, such correlations do not by themselves establish the tissue of origin, directionality, or causal transfer of ncRNAs between organs. Accordingly, throughout this section, EV-associated ncRNAs are discussed only as candidate mediators. In vivo models of Fontan palliation offer an intriguing opportunity to study heart-liver interactions. Diverse animal models, ranging from acute hemodynamic studies to chronic models designed to assess long-term complications, are being developed across multiple species to better understand Fontan pathophysiology, as detailed in by Jalal et al. [298]. For studying progressive conditions like fibrosis, chronic models are essential. For example, Van Puyvelde et al. developed an extracardiac Y-graft Fontan in sheep that achieved survival up to 21 weeks, providing a platform for preclinical investigation of long-term organ changes [299]. Earlier, Higashiyama et al. in a canine model, demonstrated how controlled increases in inferior vena cava pressure directly impacted liver hemodynamics, laying groundwork for understanding FALD [300]. As hypothesised, EVs may play a role in mediating cardio-hepatic interactions. A pilot study by the Davis lab analysed circulating EVs in a sheep Fontan model [301]. They identified distinct miRNA expression profiles, between pre- and post-Fontan EVs. Metascape analysis revealed that post-Fontan EV miRNAs target liver-specific cells and are associated with senescence and cell death, suggesting a potential role in Fontan-associated liver dysfunction [301]. However, the study does not demonstrate the cellular origin of the EVs, their trafficking to the liver, or the functional delivery of ncRNAs to hepatic cells. Later the same group also examined transcriptomic changes in patients’ circulating EVs during the perioperative period of Fontan surgery and found that post-operative EVs carried pro-angiogenic cargo that enhanced endothelial cell migration. They identified MAPK6, GLE1, miR-340-5p, and miR-199b-5p as mediators of this response. The study also showed a shift in EV tissue origin after surgery, with reduced cardiac and a slight increase in hepatic fractions, highlighting the potential of EV RNAs as candidate diagnostic and prognostic markers for Fontan-associated complications [302]. These findings are therefore most appropriately interpreted as evidence that circulating EV cargo changes during Fontan surgery, rather than as definitive proof of heart-to-liver or liver-to-heart ncRNA transfer. Another intriguing study reported that circulating exosomal miR-122, traditionally considered liver-specific, correlated with cardiac dysfunction in obesity-related cardiomyopathy. Liver-specific blockade of miR-122 attenuated abnormal cardiac remodelling [303]. Similarly, serum-derived exosomal miR-122 is associated with cardiovascular risk in NAFLD, supporting its status as a candidate circulating marker of cardio-hepatic disease [304]. Although these findings suggest a possible liver-heart link involving miR-122, they do not prove that exosomal miR-122 is causally transferred from liver to heart. CF-derived exosomes enriched in miR-21-3p induce hypertrophic signalling in CMs [305], while in the liver, miR-21-3p/miR-21-5p in exosomes promotes HepG2 proliferation [306], and weakens immune defence [307]. These findings suggest that miR-21-containing EVs are biologically active in fibrotic and inflammatory pathways in both organs, making them candidate molecules for future mechanistic studies of cardio-hepatic communication. Mesenchymal stromal cell (MSC)-derived exosomes are emerging as a promising tool for regenerative therapies [308], [309], [310], [311]. Ischemic preconditioning of MSCs induces the release of miR-22-enriched exosomes, which reduce CM apoptosis and attenuates cardiac fibrosis in vivo [308]. Hepatocyte-derived exosomal miR-22 mitigates fatty liver by inhibiting pro-lipogenic pathways [312]. Similarly, MSC-derived exosomes carrying miR-29b or miR-29a attenuate cardiac [309] and hepatic [310] fibrosis, respectively. Together, these studies support the preclinical therapeutic potential and biological plausibility of EV-associated miRNAs in cardiac and hepatic repair, although direct evidence if MSC-derived exosomes could mediate endogenous heart-liver communication remains to be established. Hypoxia- and stress-responsive miRNAs, including miR-210 and miR-214 are present in cardiac exosomes and contribute to tissue repair [311], [313]. Exosomal miR-214 improves cardiac function and reduces apoptosis and fibrosis [311], while in the liver it decreases ECM expression [314]. Macrophage-derived exosomes may also contribute to organ remodelling, with miR-34a driving CMs senescence [315], and promoting liver fibrosis [316], suggesting that similar EV-associated miRNA programmes may contribute to parallel responses in both organs. Another example of an exosome-associated miRNA is miR-223, which is abundantly expressed in immune cells, particularly neutrophils and macrophages [317]. miR-223 is an important regulator of inflammatory responses, significantly downregulated in patients with NASH [318], yet markedly increased in the border zone of infarcted myocardium [319]. In NASH, neutrophil infiltration around lipotoxic hepatocytes contributes to liver injury and inflammation [320]. Neutrophil-derived miR-223 has been shown to attenuate liver fibrosis by suppressing macrophage activation [321], [322]. Selective uptake of neutrophil-derived EVs by hepatocytes is mediated by APOE, which is upregulated in neutrophils after fatty acid treatment, and by the low-density lipoprotein receptor (LDLR) on hepatocytes [321], [323]. In addition, IL-6-activated macrophages and neutrophils can transfer miR-223 to hepatocytes via EVs, thereby reducing hepatic inflammation and fibrosis [324]. Moreover, upregulation of miR-223-3p in macrophages confers atheroprotective effects by reducing foam cell formation, limiting lipid deposition, and suppressing pro-inflammatory cytokine release [325]. Interestingly, miR-223 has also been implicated in interorgan communication between the small intestine and cardiac fibroblasts: intestinal EV-derived miR-223 was reported to promote cardiac fibrosis during aging by suppressing RASA1 and KLF15 in mouse cardiac fibroblasts [326]. Conversely, in myocardial infarction models, exosomes loaded with miR-223 improved cardiac function, at least in part by inhibiting inflammation [327]. Thus, miR-223 represents an important EV-associated miRNA potentially linking immune-cell activity with tissue fibrosis and inter-organ communication. Beyond miRNAs, lncRNAs may contribute to exosome-mediated signalling, although direct evidence for their inter-organ transfer and functional impact remains largely unestablished [328], [329], [330], [331]. In the heart, NEAT1 from hypoxic CMs exosomes promote CF survival and fibrosis [328], while macrophage-derived MALAT1 enhances autophagy in H9C2 cells [329]. Similarly, in the liver, these ncRNAs carried by exosomes promote liver fibrosis by activating HSC and increasing collagen production [330], [331]. As with miRNAs, these lncRNA studies provide mechanistic evidence within specific experimental systems, but direct evidence if EV-lncRNAs can mediate cardio-hepatic crosstalk in vivo or in humans is not yet present. An additional limitation is that EVs and exosome studies remain highly method-dependent [332], [333], [334]. Differences in EV isolation, purification and characterisation can influence the detected ncRNA cargo and make comparisons across studies difficult (Fig. 3) [299], [300]. Moreover, circulating ncRNAs may be present not only in EVs but also in non-vesicular forms, including protein-bound or lipoprotein-associated fractions [301]. Therefore, without rigorous EV characterisation and tissue-tracing approaches, it remains difficult to determine whether a specific circulating ncRNA is truly exosome-associated, which organ released it, and whether it functionally reaches a distant target organ. Collectively, these findings suggest that systemic stress and metabolic dysregulation may induce shared ncRNA programmes in the heart and liver, with EVs representing one plausible but not yet proven mechanism of inter-organ communication [4], [296], [297]. Circulating ncRNAs may therefore reflect global fibrotic burden rather than isolated organ injury [9]. Future studies using standardised EV isolation protocols, tissue-specific labelling, longitudinal human cohorts, and functional validation are required to determine whether EV-associated ncRNAs are causal mediators of cardio-hepatic crosstalk. This may enable the development of integrated biomarkers and therapeutic strategies targeting shared ncRNAs or EV-mediated signalling pathways in cardio-hepatic disease [308], [309], [310]. Conclusion and future clinical perspectives A subset of ncRNAs has emerged as potential mechanistic regulators of fibrosis in both the heart and the liver, including miR-21 [102], [105], miR-22 [110], [115], the miR-29 family [224], [319], miR-34a [121], [233], miR-122 [124], [125], miR-133a [128], [131], miR-210 [251], [255], miR-214 [136], [141], H19 [160], [163], MEG3 [168], [172], MALAT1 [173], [176], NEAT1 [178], [281] and circHIPK3 [184]. These ncRNAs may exhibit convergent (either pro-fibrotic or anti-fibrotic in both organs), divergent (opposite directions, i.e., pro-fibrotic in one organ but anti-fibrotic in the other, Fig. 7) and mixed roles (Fig. 5). However, most evidence comes from preclinical models and single mechanistic studies, thus functional validation is limited. Context, cell-type, and disease stage further complicate interpretation. Moreover, the catalogue of ncRNAs linked to both heart and liver disease remains incomplete, reflecting the field’s early stage. Emerging evidence suggests, miR-27b and miR-497, are linked to fibrosis in both heart [10] and liver [335], [336], while circRNAs like circPVT1 promote cardiac fibrosis but are poorly characterised in liver disease [337], [338], [339]. Table 6 shows key characteristics of discussed ncRNAs, detailing their mechanistic roles in fibrosis, consistency of circulating levels, associations with fibrotic markers in the heart and liver, and their current clinical development stages. Table 6ncRNAs: mechanistic roles, circulating levels, and associations with fibrosis.ncRNAMechanistic role consensus in fibrosisCirculating level consensus across studiesFibrosis predictorСlinical development stageHeartLiverStudy (ClinicalTrials.gov ID)miR-21profibrotic (convergent)up (convergent)++NCT07201363 - Recruiting patientsmiR-22mixedup (convergent)N/AN/Apreclinical onlymiR-29mixedmixed++preclinical onlymiR-34aprofibrotic (convergent)up (convergent)N/AN/Apreclinical onlymiR-122mixedmixedN/AN/ANCT06432582 - Recruiting patientsmiR-133aantifibrotic (convergent)up (convergent)N/AN/ANCT05692752 - completedmiR-210mixedup (convergent)N/A-preclinical onlymiR-214mixedup (convergent)N/A+preclinical onlyH19profibrotic (convergent)up (convergent)N/A+preclinical onlyMEG3divergent (profibrotic - liver, antifibrotic - heart)divergent (up - liver, down - heart)N/A+preclinical onlyMALAT1profibrotic (convergent)up (convergent)N/A+preclinical onlyNEAT1profibrotic (convergent)up (convergent)N/AN/Apreclinical onlycircHIPK3profibrotic (convergent)up (convergent)N/A+preclinical only ncRNAs: mechanistic roles, circulating levels, and associations with fibrosis. Recent studies support bidirectional heart-liver communication. Liver-derived proteins (LIFR [46], FXI [44], FGF21 [45]) may influence cardiac function, while heart-derived factors, like sPLA2 [40], may influence hepatic function. These findings primarily relate to protein-based communication and suggest that endocrine and paracrine signalling between the two organs might contribute to fibrotic, inflammatory, and metabolic alterations [4], [40], [44], [45], [46]. In this broader context, circulating ncRNAs, frequently associated with EVs, are hypothesised to be plausible contributors to inter-organ communication, including heart-liver signalling; however, direct mechanistic evidence for this remains largely unproven. For example, liver-enriched miR-122 has been linked to cardiac pathology [303], suggesting a potential association; however, direct mechanistic proof of heart–liver ncRNA transfer remains limited, and correlations alone do not establish causality or tissue of origin. Studies in related organ axes provide conceptual support: in the cardio-renal axis, exosomal miR-26a attenuated cardiac fibrosis in mouse models of chronic kidney disease [340], while its decrease is linked to early renal injury [341]. Diagnostic, AI, and multiplexed biomarker panels The ncRNAs’ stability in blood and ability to reflect fibrotic activity offer a promising avenue for early, non-invasive detection of remodelling [9]. This may be potentially relevant in chronic cardio-hepatic stress, such as Fontan circulation or liver diseases with secondary cardiac involvement, where conventional biomarkers lack sensitivity [5], [67], [227], [269]. However, a major constraint of ncRNA-based diagnostics is the lack of organ specificity. Many ncRNAs implicated in heart and liver fibrosis are also involved in malignancy, neurodegenerative disorders, immune-mediated disorders and fibrosis of tissues beyond the heart and liver (e.g. lung and kidney) [219], [342]. To address this challenge, a multiplexed biomarker approach is a highly anticipated strategy. Combining several ncRNAs, either among themselves or alongside established protein biomarkers, has shown promise in improving diagnostic performance in preliminary studies [343], [344]. Platforms such as nanopore-based DNA-barcoded sequencing by Oxford Nanopore Technologies [230] and magnetic bead assays developed by R&D Systems [345] provide scalable solutions for simultaneous quantification of ncRNAs and protein biomarkers and offer the potential to generate patient-specific biomarker “fingerprints” reflecting heterogeneity of disease. Nonetheless, these multi-marker panels also require extensive validation in large, diverse patient cohorts, with careful consideration of confounding factors, before clinical implementation. Artificial intelligence (AI) is increasingly emerging as a powerful complement to biomarker research due to its ability to work with complex high-dimensional and integrate multi-modal datasets [346]. A compelling example of how AI enables the analysis of large datasets is the work of Yap et al. They applied permutation importance, a feature selection method, to identify a diagnostic signature comprising 9 RNAs out of 114,602 exosomal RNAs. This signature was derived from a dataset of 118 HCC patients and 112 healthy individuals, and demonstrated the ability to accurately predict HCC [347]. Xia et al. applied machine-learning methods like Monte Carlo feature selection to determine key blood lncRNAs in HCC. Monte Carlo feature selection tests whether the expression of a given lncRNA can serve as a reliable marker to distinguish sample from a healthy person, a hepatitis B patient, a cirrhosis patient, or an HCC patient [348]. Implementation of a Bayesian network model revealed that the circulating levels of miR-122, miR-155, and miR-21 were progressively elevated with increasing fibrotic severity in patients with hepatitis C [349]. An ML-based classifier integrating plasma miR-15a/16–1 levels with AFP improved the detection of HBV-related HCC compared with AFP alone [350]. In the cardiac field, Pisklova et al. used Monte Carlo simulation and machine learning algorithms to analyse transcriptomic data and identify genes capable of distinguishing patients with hypertrophic cardiomyopathy from controls. Subsequent miRNA-gene interaction analysis revealed miR-124-3p as a potential mechanistic regulator of the largest subset of key genes [351]. By leveraging AI to analyse large transcriptomic datasets, researchers can identify ncRNAs associated with fibrosis and uncover potential regulatory networks involved in disease progression. Such approaches may help discover novel biomarkers for diagnosis, prognosis, and patient stratification, as well as reveal potential therapeutic targets. Therapeutic opportunities and translational challenges Therapeutically, shared ncRNA-regulated pathways across the heart and liver opens opportunities for dual-organ intervention. Strategies under investigation include miRNA mimics or antagomiRs [352], [353], lncRNA-targeted silencing using antisense oligonucleotides or CRISPR-based approaches [354], [355], and exosome-mediated delivery of protective ncRNAs [289]. MiRNA mimics are synthesised double-stranded RNA molecules imitating endogenous miRNA, while antagomiRs are single-stranded RNA molecules designed to target specific miRNAs and prevent their interactions with target genes [356]. For instance, miRNA-21 silencing with miRNA-21- antisense oligonucleotides (ASO) increased the expression of its target gene HBP and improved steatosis [357]. miR-221 regulates CD44 expression in HCC, and treatment with miR-221-ASO downregulates CD44 protein levels [358]. Conversely, synthetic mimics of anti-fibrotic miRNAs can be introduced to restore their levels, thereby boosting protective pathways and suppressing fibrosis. Examples include treatment with miR-378a-3p mimic via nanoparticles reducing liver fibrosis in vivo murine model [359], and in vivo treatment with miRNA mimics attenuated myocardial ischemia and reperfusion injury [360], decreased myocardial injury after MI [361]. Modulation of lncRNA expression can also be achieved using ASOs, small interfering RNAs (siRNAs), and aptamers [362]. siRNAs are double-stranded RNA molecules that, upon incorporation into the RISC, guide Argonaute-mediated cleavage and degradation of the target lncRNA [363]. For example, silencing of lncRNA Wisper with ASO attenuated MI-induced fibrosis and cardiac dysfunction in mice [364]. One of the principal translational challenges in ncRNA-based therapeutics is the targeted delivery of therapeutic molecules to specific organs [356]. Synthesised miRNA mimics and antagomiRs lack intrinsic affinity for cell-surface receptors, necessitating the use of specialised delivery systems. Currently under investigation are lipid nanoparticles (LNP), polymeric carriers, and viral vectors [365], [366]. For example, LNPs have demonstrated particular promise for hepatic delivery [367], while adeno-associated virus serotype 9 have shown efficiency in myocardial targeting in mice [368]. A further fundamental challenge concerns the biological specificity of ncRNA-based therapies. Given that a single miRNA can simultaneously regulate hundreds of target genes across multiple pathways, the introduction of miRNA mimics or inhibitors carries significant risk of unintended off-target effects, potentially disrupting non-pathological gene regulatory networks [356]. Additionally, the introduction of viral vectors or exogenous synthetic nucleic acids can trigger innate and adaptive immune responses, leading to inflammation, cytotoxicity, and compromised therapeutic efficacy [356], [369]. Despite substantial progress in identifying ncRNAs associated with fibrosis, several challenges must be addressed before these findings can be translated into clinical applications. Comprehensive integration of genomic, transcriptomic, proteomic, and metabolomic data will be required to fully map ncRNA-mediated regulatory networks implicated in disease pathogenesis [80], [223]. Multi-omics-guided prioritisation represents a powerful and systematic platform for identifying therapeutically relevant ncRNA candidates. Zhang et al. integrated transcriptomic with proteomics and 69 concordantly altered molecules, among 1224 differentially expressed genes, 302 differentially expressed proteins, linked to metabolism, detoxification, and ECM remodelling in liver fibrosis in mice [370]. Yang et al. profiled liver and plasma samples from HBV, ALD, and NAFLD patients, identifying shared fibrotic signatures and 132 circulating protein biomarkers that predicted advanced fibrosis better than FIB-4 [371]. Untargeted proteomics and metabolomics of HFpEF plasma samples revealed distinct inflammatory and immune response pathways in HFpEF [372]. Integrating ncRNA expression profiles with metabolite–protein interaction networks can help prioritise functionally important ncRNAs by linking them to dysregulated metabolic and protein pathways. This network-based approach distinguishes central pathological regulators from peripheral or less relevant ncRNA changes, thereby improving the identification of candidate biomarkers and therapeutic targets in hepatic and heart fibrosis [373]. However, while multi-omics approaches can prioritise candidate ncRNAs and regulatory networks, their clinical relevance will depend on validation across time, disease stage, and independent patient populations. Longitudinal profiling in independent cohorts will be essential to identify early ncRNA alterations that precede overt tissue remodelling, as ncRNA expression is highly dynamic and cell-state-dependent [374]. Rigorous functional validation in relevant disease models remains critical to establish causality and therapeutic efficacy, as correlative expression changes alone are insufficient to define ncRNA function within complex fibrotic network. Furthermore, context-dependent effects, where a single ncRNA may exert pro-fibrotic effects in one tissue and protective roles in another, must be carefully delineated to avoid unintended consequences of systemic targeting [375]. In summary, ncRNAs represent both mechanistic drivers and promising biomarker candidates currently under investigation in multisystem fibrotic disease. Nonetheless, significant challenges related to cohort size, validation, confounding factors, and organ specificity must be addressed for their successful clinical translation. Integrating ncRNA biology with multiplexed detection technologies and personalised analytical frameworks might enable a transition from traditional organ-centric paradigms, toward coordinated diagnosis and management of cardio-hepatic fibrosis. CRediT authorship contribution statement Giovanni Battista Luciani: Writing – review & editing, Conceptualization. Julia Gorelik: Writing – review & editing, Supervision, Conceptualization. Paolo Ferrero: Writing – review & editing, Writing – original draft, Conceptualization. Pier Paolo Bassareo: Writing – review & editing, Conceptualization. Christine Voellenkle: Writing – review & editing, Writing – original draft, Supervision, Conceptualization. Germana Zaccagnini: Writing – review & editing, Conceptualization. Veronika Boichenko: Writing – review & editing, Writing – original draft, Visualization, Conceptualization. Svetlana Frolova: Writing – review & editing, Writing – original draft, Visualization, Conceptualization. Fabio Martelli: Writing – review & editing, Supervision, Conceptualization. Declaration of Generative AI and AI-assisted technologies in the writing process After writing, some portions of the text were edited using ChatGPT-5.1 to increase text clarity and synthesis. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. Funding V.B. is supported by a PhD fellowship from the University of Verona, Italy. S.F. is supported by the PhD School in Molecular and Cellular Biology, University of Milan, Italy. This study was also supported by BHF to J.G., Grant no. RG/F/22/110081. F.M. is partially supported by Ricerca Corrente funding from Italian Ministry of Health to IRCCS Policlinico San Donato (#1.07.128; #1.07.127). F.M. is also supported by the Italian Ministry of Health (POS-T4 CAL.HUB.RIA T4-AN-09), and by the European Union (Next Generation EU-NRRP M6C2 Inv. 2.1 PNRR-MAD 2022-12375790 and PNRR-MCNT2-2023-12377983, and Romania’s PNRR-III-C9-2022-I8, CF 186/24.11.2022, contr. 760062/23.05.2023). 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 We thank Laura Perfetti, Scientific Directorate at IRCCS Policlinico San Donato, for proofreading the manuscript and for improving its clarity, readability, and language. Figures were created with Biorender assessed on 28th December 2025. Graphical abstract - https://BioRender.com/enqvu9m; Fig. 1 - https://BioRender.com/chkhaol; Fig. 4 - https://BioRender.com/iyqysv6; Fig. 5 - https://BioRender.com/i8fsbz3; Fig. 6 - https://BioRender.com/815z7hr; Fig. 7 - https://BioRender.com/y8qsqb1. Contributor Information Julia Gorelik, Email: j.gorelik@imperial.ac.uk. Fabio Martelli, Email: fabio.martelli@grupposandonato.it. Data availability No data was used for the research described in the article. References Associated Data Data Availability Statement No data was used for the research described in the article.