Role of Exosomal miRNAs in Heart Failure

Heart failure is the terminal outcome of the majority of cardiovascular diseases, which lacks specific diagnostic biomarkers and therapeutic targets. It contributes to most of cardiovascular hospitalizations and death despite of the current therapy. Therefore, it is important to explore potential molecules improving the diagnosis and treatment of heart failure. MicroRNAs (miRNAs) are small non-coding RNAs that have been reported to be involved in regulating processes of heart failure. After the discovery of miRNAs in exosomes, the subcellular distribution analysis of miRNAs is raising researchers' attention. Growing evidence demonstrates that exosomal miRNAs may be promising diagnostic and therapeutic molecules for heart failure. This review summarizes the role of exosomal miRNAs in heart failure in the prospect of molecular and clinical researches.


INTRODUCTION
Heart failure is the terminal stage of various CVDs, with sustaining high morbidity and mortality contributing most to cardiac death (1). Despite of the current therapy strategy, the mortality and the re-hospitalization of heart failure are still high and no specific drugs are found to reverse heart failure once initiated. Therefore, exploring the pathogenesis of heart failure and labeling molecules closely related to the occurrence and development of heart failure are quite important, since these actions can post significant impact on early detection and treatment of the disease which can be prevented from adverse events.
MicroRNAs (miRNAs) are small non-coding RNAs that have been discovered for over two decades. MiRNAs mainly function as negative regulators of post-transcription of gene expressions through hybridization of 3 ′ untranslated regions (3 ′ UTRs) or open reading frames (ORFs) of the targeted mRNAs. Although they don't encode any protein, they are widely involved in the modulation of various genes at post-transcription level and thereby impacting various biological processes such as autophagy, oxidative stress and so on (2)(3)(4)(5). Exosomes are bilayer membrane extracellular vesicles (EVs) carrying lipids, proteins and nucleic acids. They can be secreted by almost all eukaryotic cells through the fusion of multivesicular bodies and plasma membranes. They are released to extracellular matrix and responsible for carrying biological regulatory information [proteins, lipids, DNAs, and RNAs, especially miRNAs (6)] to the target cells in both surrounding and distant sites and organs, serving as paracrine signaling mediating intercellular communications. Exosomes transport intercellular information via four mechanisms: (1) Exosomal membrane proteins bind to the ligand-receptors in membranes of the target cells directly, activating certain signaling pathways; (2) Exosomal membrane proteins are splintered into pieces, releasing soluble ligands that bind to receptors of recipient cells; (3) Exosomes membranes directly fuse with the membranes of the target cells; (4) Internalization of exosomes by endocytic processes such as phagocytosis and macropinocytosis (7).
MiRNAs distribute widely in cellular components, and recent studies revealed that miRNAs also exist in exosomes (exosomal miRNAs, exo-miRNAs) (8). Furthermore, recent evidences revealed that miRNAs can be isolated easily and steadily from many kinds of tissues and body fluids including circulation, urine and saliva. In circulation, the detective miRNAs are mainly in the forms of exosomal miRNAs. The miRNAs are steady due to the protection of exosomes, whereas miRNAs in free forms will be degraded by nuclease (9). The discovery of exosomal miRNAs especially in circulation broadens functions of miRNAs to intercellular paracrine signaling mediators and disease biomarkers. In heart failure, mounting evidences have suggested that miRNAs participate in its pathogenesis and progression. However, attention focused on the component analysis of miRNAs and the related mechanisms in heart failure are still lacking. This review summarizes the recent evidences and progress of exosomal miRNAs in heart failure.

EXOSOMAL miRNAs: ORIGIN, FUNCTION AND MECHANISM IN CVDs
Exosomal miRNAs are important biological products in exosomes, which are widely involved in information transmission between cells. It is worth noting that miRNAs do not enter exosomes randomly, but specific miRNAs are selectively entered into exosomes. Mature miRNA enters exosomes through four ways: (1) the miRISC related pathway; (2) sphyngomyelinase 2 dependent pathway; (3) heterogeneous nuclear ribonucleoproteins dependent pathway; (4) 3 ′ miRNA sequence-dependent pathway. These are important ways for exosomal miRNAs to achieve directional information exchange between cells. Exosomal miRNAs participate in a variety of basic cell functions through paracrine or endocrine mechanisms, including cell proliferation, apoptosis, cytokine production, immune regulation and metastasis (10,11). Under pathophysiological conditions, the content of specific exosomal miRNAs will also change (12). This makes evaluating levels of exosomal miRNAs a way for us to diagnose (13)(14)(15)(16)(17) and treat diseases (18). In the cardiovascular field, exosomal miRNAs normally play the role as cardiovascular cells communicators. Almost all kinds of cardiovascular cells are proved to release and receive exosomal miRNAs, indicating their potential regulatory roles in CVDs. Cardiomyocytes and fibroblasts, the most prevalent cell types in the heart, are the main source of exosomal miRNAs. Other than that, ECs and macrophages are also proved to release exosomal miRNAs in different CVDs (19). Depending on the source of exosomal miRNAs, target cells and the disease settings, these exosomal miRNAs act as different regulators on various molecular pathways in different CVDs [such as myocardial infarction, cardiomyopathy, arrhythmia, atherosclerosis and so on (20), Table 1], and the mechanisms involved are various. However, none of the studies successfully illustrate how exosomal miRNAs fulfill their specificity and further illustrated the way specific miRNA incorporate into exosomes. They use exosomes as non-selective transporters rather tissue/cell specific transmitters. Based on the above, we believe that there is still a long way for the research of exosomal miRNAs in CVDs.

EXOSOMAL miRNAs IN HEART FAILURE
Heart failure is the terminal stage of almost all cardiovascular disorders mentioned above. It is characterized by a series of pathological processes including cardiomyocyte hypertrophy, cardiac fibrosis, impaired myocardial angiogenesis. Owing to the unique traits of exosomal miRNAs, investigating how they affect the pathogenesis of heart failure is meaningful.

Exosomal miRNAs in Cardiomyocyte Hypertrophy as Regulators
A series of miRNAs have been demonstrated to modulate the pathogenesis of cardiac hypertrophy which is the key mechanism of heart failure. MiR-199a targets the glycogen synthase kinase 3β (GSK3β)/mammalian target of rapamycin (mTOR) complex signaling pathway to impair cardiomyocyte autophagy and thereby enhancing cardiac hypertrophy (miR-199a impairs autophagy and induces cardiac hypertrophy through mTOR activation). In the model of heart failure, cardiacspecific overexpression of miR-221 significantly deteriorates cardiac function and promotes heart failure and mTOR axisrelated autophagy (29). However, few of the studies above have analyzed the component and the origin of the miRNAs involved in the regulation of cardiac hypertrophy. Furthermore, how miRNAs in non-cardiomyocytes exert their anti-hypertrophic effects on cardiomyocytes remains unexplained.
Recently, more and more evidences indicate that exosomal miRNAs might illustrate the issues above well Figure 1. Under pro-hypertrophic stresses, exosomal miRNAs from various kinds of cardiovascular cells target at cardiomyocytes, initiating or inhibiting cardiac hypertrophy. Based on the origin of exosomal miRNAs, there are basically three kinds of mechanisms involved: (1) Communication among cardiomyocytes: In MI model, exo-miRNA-133a level was significantly downregulated in cardiomyocytes in the infarct and peri-infarct areas. Exo-miR-133a released from ischemic cardiomyocytes  (19). MiR-21, a proven regulator of fibroblasts biology, is increased in failing hearts, implicating its potential role in heart failure. However, overexpressing or inhibiting miR-21 in cardiomyocytes doesn't influence the hypertrophic phenotype, raising the question that the origin or distribution of miR-21 might be involved. Further evidences suggest that miR-21 expressions are extremely high in exosomes and miR-21 expressions are higher in donor fibroblasts, indicating that miR-21 is packed into exosomes. Furthermore, exo-miR-21 derived from fibroblast is proved to be absorbed into cardiomyocytes and subsequently decreases SORBS2 and PDLIM5 in cardiomyocytes, ultimately inducing cardiac hypertrophy. Thus, exo-miR-21, rather than miR-21, is a key mediator transferring pro-hypertrophy information from cardiac fibroblasts to cardiomyocytes (19). MiR-217 was also proved to be a potent marker of chronic heart failure, evidenced by increased expressions in heart failure and its overexpression pressure overload-induced cardiac hypertrophy. Further component analysis proved that cardiomyocyte-derived exosomes containing miR-217 enhanced proliferation of fibroblasts in vitro, suggesting that exosomal miR-217 was not only a potent marker but a promising therapeutic target for chronic heart failure (21). (3) Communications between other cardiovascular cell types and cardiomyocytes: Exosomes enriched with miR-155 which were derived from macrophages promoted cardiac hypertrophy and fibrosis by activating the prohypertrophic pathway FOXO3a in uremic mice (25). In acute MI, exosomal miRNAs also acted in cell-cell communication between cardiomyocytes and local or remote stem cells. For example, exosomes enriched with miR-451 that derived from cardiac progenitor cells protected cardiomyocytes from oxidative stress by targeting GATA4, thereby maintaining cardiac function. Exo-miR-146a extracted from ECs were absorbed by cardiomyocytes and subsequently targeted on Erbb4, Notch2, Irak1, mediating the progression of heart failure through metabolic pathways (23). (4) Communications between other cell types in remote organs and cardiomyocytes: A recent research proved that PPAR-γ activation in adipocytes from adipose tissues increased the secretion of miRNA-200-contained exosomes which promoted cardiomyocyte hypertrophy through the mTOR pathway (26). Exosomal miRNAs are involved in the regulation of cardiac hypertrophy due to various etiology. However, based on the current evidences, the defined role of exosomal miRNAs in cardiac hypertrophy differs across disease settings and the sources, not to mention the miscellaneous mechanisms involved. We still lack further research illustrating how exosomes ensure the specificity for certain miRNAs targeting at specific cell types.

Exosomal miRNAs in Cardiac Fibrosis as Regulators
Cardiac fibrosis, characterized by ECM degradation and accumulation of collagen due to proliferation of cardiac fibroblasts, is another important mechanism of heart failure. MiRNAs have been reported to regulate cardiac fibrosis. For example, miR-19a-3p and miR-19b-3p targeted TGF-β to inhibit epithelial-mesenchymal transition, ECM production and the invasion of human cardiac fibroblasts, thereby exhibiting the anti-fibroblast effect (34). More lights are shed on the regulation of cardiac fibrosis by exosomal miRNAs. Exosomal miR-294 from embryonic stem cells were proved to inhibit fibrosis after MI, thereby preventing MI-induced heart failure. Elevated exo-miR-29a in marginal ischemic area transferred the intercellular information among cardiomyocytes, mediating the antifibrotic effect and preventing ventricular dysfunction and heart failure (35). MiR-208a was upregulated in cardiomyocytes and cardiomyocyte-derived exosomes in cardiac fibrosis (36).
Mechanically, exosomal miRNAs from diverse sources participate in the regulation of different parts of cardiac fibrosis Figure 1. Macrophage-derived exo-miR-155 inhibited the proliferation of cardiac fibroblasts after MI by inhibiting the SOS-1, a critical modulator of RAS activation in mice (37). Cardiomyocyte-derived exosomes containing miR-208a contributed to increased fibroblast proliferation and differentiation into myofibroblasts and thereby accelerating the progression of cardiac fibrosis (36). In type 2 diabetes mice, exercise significantly increased exosomes derived from cardiomyocytes which contain abundant miR-29b and miR-455, thereby decreasing MMP9 and ECM degradation and resulting in the amelioration of cardiac fibrosis (38). In diabetic cardiomyopathy, exosomes derived from diabetic cardiomyocytes negatively affected proliferation and migration of ECs by transferring miRNA-320 and subsequently inhibited heart failure. This process was reversed by adding GW-4869, an exosome inhibitor (22). Exosomal miR-208 derived from cardiomyocytes inhibited the expression of Dyrk and phosphorylated NFAT to promote its nuclear export. Nuclear NFAT triggered fibrosis by inducing the expression of CTGF, Col1a1, Col3a1, and a-SMA in cardiac fibroblasts (36). The evidences above have proved the fact that miRNAs in exosomes are important regulators of cardiac fibrosis and heart failure due to different etiologies. Injection of exo-miR-26a into tibialis anterior (TA) muscle significantly increased the expression of miR-26a in the heart and subsequently ameliorated cardiac fibrosis lesions. Furthermore, exo-miR-26a significantly blunted FOXO1 activation and inhibited cardiac fibrosis in the heart of mice with chronic kidney disease (39). This study demonstrated the strength of miRNAs encapsulated by exosomes for its tissue specificity, however, how exosomes guarantee this specificity needs illustrating. Furthermore, whether this specificity was due to the encapsulation of exosomes or miRNAs themselves are yet unknown.

Exosomal miRNAs in Impaired Myocardial Angiogenesis as Regulators
Besides of the pathologic changes in heart tissue, aberrant angiogenesis contributes to the acceleration from compensated cardiac hypertrophy to heart failure due to decreasing capillary angiogenesis and cardiomyocyte energy supply (40). ECs proliferation is the key factor in this process (41). Exosomal miRNAs have been proved to regulate angiogenesis in other diseases Figure 1, but how they affect angiogenesis in heart failure is not fully and directly illustrated.
Exo-miR-21-5-p from healthy hearts inhibited PTEN and BCL2 in cardiomyocytes and ECs and activated the AKT and VEGF pathways to promote cardiomyocyte proliferation and promote angiogenesis (42). Exosomes containing miR-939 from hearts in patients with myocardial ischemia significantly promoted angiogenesis through iNOS-NO pathway (43). However, the source of exosomal miRNAs in the researches above might need further study. In the subsequent studies, they perfectly demonstrated the sources, target cells and the mechanisms of exosomal miRNAs in the regulation of angiogenesis. Wang et al. discovered that exosomes isolated from diabetic cardiomyocytes significantly reduced angiogenesis in the diabetic myocardium via the exosomal transfer of miR-320 into ECs and subsequent decrease of IGF-1, Ets and Hsp20 (22). In MI, M1-like macrophages released a large number of proinflammatory exosomes, transferring miR-155 to ECs to inhibit angiogenesis by down-regulating Rac family small GTPase 1 and p21 activates kinase 2 (44). Interestingly, we discovered the same miRNA might exert completely opposite regulatory effects in angiogenesis in the process of heart failure. MiR-146a in the exosomes derived from ECs significantly decreased ECs proliferation and inhibited microvascular regeneration and impaired systolic/diastolic function in cardiomyopathy (23). However, exosomal miR-146a derived from cardiosphere cells promoted myocardial angiogenesis and significantly improved heart failure after acute MI (45). Some scholars believe that these contradictions may be related to different periods of heart failure. In the early stage, cardiac exosomes transmitted cardioprotective miRNAs, and in the terminal stages, exosomal miRNAs may play a damaging role (46). In addition, we also assume that the seemingly opposite results might be attributed to the origin of miRNA and different etiologies of heart failure. Further mechanistic research demonstrating this contradiction is required.

APPLICATION OF EXOSOMAL miRNAs IN HEART FAILURE: BIOMARKERS AND THERAPEUTIC TARGETS
The exploration of new biomarkers of heart failure is critical for the diagnosis and treatment of heart failure. BNP and NT-proBNP are widely used in the clinical diagnosis of heart failure. Some newly found biomarkers such as ST2 and NGAL also become to be used in clinical. However, in the case of combination with some other pathological conditions, such as pulmonary embolism and renal dysfunction, the specificity of the biomarkers mentioned above could be compromised.
A variety of miRNAs in circulation are involved in pathogenesis or protection of heart failure in a dynamic and stage-specific way and therefore they are potential biomarkers of heart failure. Circulating miR-23a, miR-23b, miR-24, miR-195, and miR-214 were found increased and circulating miR-423-5p, miR-320, and miR-22 were found decreased in chronic heart failure (47,48). As for acute heart failure (AHF), miR-423-5p was proved to differentially express between heart failure patients and controls, which might serve as a biomarker distinguishing AHF and other causes of dyspnea (49). MiR-210 in plasma was also proved to be elevated in heart failure patients with NYHA III-IV compared with those with NYHA II and controls (50). Other differential expression circulating miRNAs in AHF includes down-regulated miR-103, miR-142-3p, miR-30b, miR-342-3p and up-regulated miR-499 (51,52). Furthermore, miRNAs were also proved to have potential ability to evaluate prognosis. Decreased levels of miR-18a-5p and miR-652-3p during hospitalization in heart failure patients were associated with a high predicted mortality within 180 days (53). Up-regulation of miR-192 was proved to be a prognostic marker predicting survival in ischemic heart failure with p53-mediated apoptosis involved (54).
However, none of the studies above analyzed the circulating components, especially exosome fraction and non-exosome supernatant. Wang et al. discovered that miR-425 and miR-744 expressions were altered in plasma exosome samples in heart failure patients. A further functional study proved that they negatively regulated cardiac fibrosis by suppressing TGFβ1 Frontiers in Cardiovascular Medicine | www.frontiersin.org expression, revealing their potential role in regulating and predicting heart failure (55). However, direct evidence of the regulatory role of miRNAs via exosomes was lacking in this study. Circulating miRNAs exist in the form of free miRNAs and exosome-encapsulated miRNAs. Free miRNAs in circulation are the byproduct of dead cells and their levels might be altered depending on the rates of tissue repair and age. On the contrary, exosomal miRNAs, making up 10-15% of circulating miRNAs, are more sensitive in providing stable and accurate patterns for disease diagnosis. However, the difficulty to isolate and detect circulating exosomal miRNAs limited their further application as heart failure biomarkers. As the technology develops, Turchinovich et al. firstly separated non-exosome supernatant fraction from exosome and Beg et al. constructed protocols to measure both free miRNAs and exosomal miRNAs in circulation, making it possible to proceed the biomarkers exploration focusing on exosomal miRNAs (56,57). In the study of Beg's, they found an increase of miR-146a and miR-486 (both had the poverty of anti-inflammation) in exosomes rather than those in plasma in the patients with heart failure compared to controls, indicating that exosomal miRNAs worth being further studied as heart failure biomarkers. Inspiringly, two studies supported the diagnostic value of exosomal miRNAs. Serum exo-miR-92b-5p was proved to be increased in HFrEF patients compared with controls, with a sensitivity of 71.4% and specificity of 83.3%. Moreover, exo-miR-92b-5p was inversely correlated with the LVFS and LVEF, while positively correlated with LAD, LVDD and LVSD. Serum exo-miR-92b-5p is a potential diagnostic biomarker in AHF due to DCM, the elevation of which indicated the progression of AHF (58). In another study focused on ischemic HF post AMI discovered that both serum levels and exosomal levels of miR-192, miR-194, and miR-34a were up-regulated in HF patients, confirming their diagnostic role in ischemic HF (59). However, this study did not compare the disparity between serum miRNAs and exosomal miRNAs regarding to their diagnostic value. In myxomatous mitral valve disease model of dogs, decreased exosomal miR-599 levels and increased exosomal miR-181c and miR-495 were found in chronic heart failure compared to controls, whereas no miRNA in total plasma remained statistically significant alteration if the false discovery rate was set <15%, indicating that exosomal miRNAs might be a more specific marker than total plasma miRNAs (60). The concern of the comparation of the diagnostic value between exosomal miRNAs and free miRNAs are raised consequently. Some studies supported that exosomal miRNAs exhibit more diagnostic and prognostic value than those in free form. For example, in stable coronary artery heart disease, miR-126 and miR-199a in exosomes were proved to be related with cardiovascular events rather than free miR-126 and miR-199a in plasma (61). However, a study opposed this opinion with the evidence that no differential expression of exosomal miRNA and free-miRNA in circulation were detected between chronic heart failure patients and controls (48). In a 30 patients' clinical study, the disparity of miRNA levels (miR-423-5p/320a/22/92b/17/532-3p/92a/30a/21/101) between patients with heart failure and controls were similar in exosomal fraction and unfractionated serum, suggesting that these miRNAs in exosome might not be ideal markers for heart failure. However, we doubt that the sample capacity in this study was small and the coverage of miRNA detected was limited (48). Further studies are desperately needed to demonstrate the diagnostic value of exosomal miRNAs in heart failure by expanding the number of patients enrolled and the miRNAs. Furthermore, the source of most of the exosomal miRNAs detected in circulation has not been further illustrated, limiting the mechanism exploration of the specific exosomal miRNAs.

STEM CELLS-DERIVED EXOSOMAL miRNAs IN HEART FAILURE: OPPORTUNITIES AND CHALLENGES
The stem cell-based therapies are becoming potential therapeutic strategies in many diseases and their influences radiate to heart failure, whereas the low efficiency of engraftment and potential tumorigenic risk of stem cell transplantation limits their prevalence. However, the emerging value of stem cellderived exosomes which might improve the efficiency of stem cell transplantation motivates the further exploration of stem cellbased therapies in heart failure. Exosomes derived from different kinds of stem cells, such as ESCs, iPSCs CPCs and MSCs, have been proved to exert no less protective properties but safer in operation than stem cell transplantation (62). As one of the major components in exosomes, several miRNAs in exosomes derived from stem cells have been identified to play important roles in heart failure. Intramyocardial delivery of exosomes enriched with miR-290-295 and miR-294 derived from ESCs significantly ameliorated fibrosis in acute MI. Exosomes from murine cardiac fibroblast-derived iPSCs, which were enriched with miR-21 and miR-210, have been proved to protect against oxidative stress and cardiac remodeling after acute MI in periinfarct regions. And the protective properties of miR-21 enriched exosomes were similar with miR-21 (63,64). CPCs treatments help to improve cardiac repair and protect against damage and exosomes in CPCs are involved. Proangiogenic miR-17 and miR-210 were up-regulated in CPC-derived exosomes under oxidative stress and thereby inhibited TGF-β induced fibrosis, providing another profound evidence that miRNAs in exosomes might be a pivotal component in the stem cell therapy of heart failure (65). Researchers found that bone marrow-derived MSC (BMSC) from mice after ischemic preconditioning could secrete miR-22 highly enriched exosomes, and in the MI mouse model, miR-22 targeted methyl CpG binding protein 2 (Mecp2) to reduce infarct size and cardiac fibrosis (66). After treating human engineered cardiac tissue (hECT) with exosomes derived from human mesenchymal stem cells (hMSCs), the level of miR-21-5p is significantly increased in hECT, while knockout of miR-21-5p in hMSCs reduced expression of pro-contraction and related calcium-treated genes in the hECT (67). These findings indicate that stem cell-derived exosomes contain a variety of miRNAs which play an important role in heart failure ( Table 2). The use of these exosomes may contribute to the development of novel treatment strategies for heart failure in the future.

PROSPECT
The types and levels of miRNAs in exosomes in the circulatory system are more closely related to heart failure compared to free miRNAs. Other than free miRNAs, their level changes may be related to the severity of heart failure. So, they can provide a more sensitive and specific diagnostic method for heart failure. However, there is evidence that no differential expression of exosomal miRNAs and free miRNAs in the circulation was detected between patients with chronic heart failure and the controls. Therefore, whether circulating exosomes miRNAs can be used as biomarkers for the diagnosis and prognosis of heart failure is still questionable. We need studies with larger number of patients and the kinds of miRNAs to determine: whether there are significant changes in the levels of circulating exosomal miRNA after heart failure, whether there is a correlation between these changes and the pathophysiological changes in heart failure, whether the sensitivity and specificity of these new diagnostic methods are better than those currently used in clinical practice. Heart failure may affect circulating exosomal miRNAs, and exosomes derived from stem cells can also play an important role in heart failure. They may serve as a new strategy for treating or preventing heart failure in the future. However, this new strategy is in its infancy. There are still several hurdles before achieving clinical application (62): First of all, the purification and separation technology of exosomes is not perfect, it takes a long time and the efficiency is low (69). Secondly, exosomes contain a large number of biologically active factors, so they may cause adverse or side effects on cardiac tissue. Thirdly, due to the heterogeneous components in exosomes, it may show a potential risk for tumor formation and a risk of adverse immune response based on donor cell characteristics. Finally, due to the complex structure of exosomes, the therapeutic effect that exosomes produced during the disease is unknown. How exosomal miRNAs fulfill its specificity is yet fully understood. Therefore, the future clinical translation of exosomes for heart failure treatment still faces major challenges. We need to clearly analyze the various components in stemderived exosomes, clarifying their physiological effects on cardiac tissue, as well as their therapeutic and side effects on human body.

AUTHOR CONTRIBUTIONS
RX, ZX, PH, and WT wrote the manuscript. YW drew the figure. JH, CL, and YD edited the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
The present study was funded by the National Natural Science Foundation of China (Nos. 81700344).