MicroRNAs in the failing heart – Novel therapeutic targets?

Abstract

Heart failure is a common and disabling disease with high mortality that carries substantial societal costs. Current therapeutic strategies are aimed at relieving symptoms, avoiding hospitalization, and prolonging life, but disease progression is ultimately inevitable. MicroRNAs (miRNAs) are short, non-coding RNA molecules with pervasive roles in eukaryotic biology, annealing to complimentary sites on target mRNAs, and repressing gene expression. The fact that miRNAs are dysregulated in many human disorders, including cardiovascular disease, and the relative ease with which endogenous miRNA expression can be altered using synthetic antisense oligos has stirred enthusiasm for these molecules as potential drug targets. The aim of this review article was to summarize the current knowledge on the roles of miRNA in the pathophysiology of heart failure as well as the use of miRNAs as therapeutic targets and diagnostic tools for the disease.

Key words::

• diagnostic biomarkers
• heart failure
• microRNA
• Therapeutic targets
• review

Introduction

Heart failure (HF) represents a common final manifestation of many cardiac and cardiovascular diseases, including ischemic heart disease, hypertension, atrial fibrillation, valvular heart disease, and cardiomyopathy. The fact that HF is both common (approximately 2% of the population in high-income countries) and disabling makes HF a massive burden on health services and societies (1,2). Although current therapies, including drugs and mechanical assist devices, can alleviate symptoms and prolong life, therapies to fully prevent progression of the disease are still lacking. Thus, there is a need for novel therapeutic strategies.

In recent years, there has been a surging interest in the non-protein-coding parts of the genome. It is believed that although protein-coding genes represent only 1% of the human genome, more than 90% of genomic bases are actively transcribed (3). The importance of these non-coding transcripts in many fundamental biological and pathophysiological processes are rapidly being revealed. One class of non-coding RNA that has been the subject of extensive research in recent years is microRNA (miRNA). These short (approximately 22 nucleotides) RNA molecules repress gene expression through base pairing with target sites on specific mRNAs, thereby facilitating degradation or translational repression(4). Since the discovery of miRNAs in 1994, these molecules have now been implicated in virtually every biological process in higher eukaryotes.

This review will focus on the role of miRNA in HF, its role in the patophysiology of the disease, as well as the potential use of miRNAs as therapeutic targets and diagnostic tools.

The role of miRNAs in the normal and failing heart

miRNAs in cardiac development and function

One of the first papers describing a role for microRNAs in cardiac biology was published in 2005. Zhao et al. showed that the cardiac and skeletal muscle-specific miRNA miR-1 is expressed during cardiogenesis and regulates the central cardiac transcription factor Hand2 (5). Overexpression of miR-1 resulted in developmental arrest, thin-walled ventricles, and HF in mouse embryos. In another study, Kwon et al. showed the importance of miR-1 in cardiac development in Drosophila where it was shown to regulate specific cardiac and somatic muscle lineages from progenitor cells through Notch signaling (6).

In another approach to understand the role of miRNAs in cardiac function and development, Chen et al. focused on Dicer, an endonuclease that is essential for processing of immature precursor miRNAs (pre-miRNAs). Cardiac-specific deletion of Dicer (and subsequent depletion of all cardiac miRNAs) resulted in progressive dilated cardiomyopathy, HF, and postnatal lethality (7). Furthermore, the hearts of Dicer mutant mice were characterized by a substantial decrease in the number of sarcomeres, decreased contractile force, and impaired cardiac conduction. Interestingly, Dicer was found to be downregulated in failing human hearts, an effect that was reversed with the introduction of a left ventricular assist device.

Using a similar approach, Rao et al. deleted another component of the miRNA processing pathway, dgcr8, which is part of the nuclear protein complex necessary for cleaving primary miRNAs into 60–80-bp pre-miRNAs (8). Mutant mice died within 2 months with a severe cardiac phenotype. In line with the findings from the Dicer knockout mice, dgcr8^−/− mice had dramatically reduced left ventricular function and impaired cardiac conduction.

miRNAs in the cardiac stress response

The increased demands on ventricular pressure and volume in the failing heart triggers cardiac stress responses, including hypertrophy, fibrosis, and rearrangement of sarcomeres (9). These pathological changes are associated with altered gene expression and the activation of fetal gene programs (10). Thum et al. compared the transcriptome in fetal hearts with explanted hearts from end-stage HF patients and found striking similarities (11). There was approximately 85% overlap in the differentially expressed miRNAs between the two tissue types and the authors show that miRNAs contribute substantially to the dysregulation of gene expression associated with HF. Prasad et al. used Ingenuity Pathway Analysis to study the effect of differentially expressed miRNAs on gene networks in failing human hearts (12). The authors identified a handful of key molecules targeted by HF-associated miRNAs that might consitute potential therapeutic targets.

One hallmark of gene reprogramming in HF is the activation of the gene β-myosin heavy chain (βMHC), which is normally expressed exclusively during embryogenesis, and the reciprocal repression of αMHC, which is expressed in the adult heart (13). Van Rooij and co-workers identified miR-208, which is situated in an intron of αMHC and is expressed exclusively in the heart, as an important regulator of the cardiac stress response(14). Deletion of the miR-208 gene rendered mice virtually resistant to various hypertrophic stimuli, due to an inability to upregulate βMHC. In a separate paper, the same group identified a set of overlapping miRNAs that were overexpressed in mouse models of cardiac hypertrophy as well as in human HF patients. One of these miRNAs, miR-195, was found to be sufficient to drive cardiac hypertrophy and cause dramatic effects on cardiac structure, function, and gene expression (15).

Gurha et al. identified miR-22 as a key player in the cardiac stress response. They showed that genetic ablation of the miR-22 gene led to an increased propensity of mice developing cardiomyopathy in response to pressure overload (16). After 4 weeks of trans-aortic constriction (TAC), mice lacking miR-22 had an increased left ventricular end-systolic dimension and impaired fractional shortening compared with wild-type animals. The authors identified the transcriptional co-factor PURB as a direct target of miR-22 and the downstream effects of PURB derepression as a possible mechanism for the cardiac phenotype. In a later study, Huang and colleagues explored the role of miR-22 in response to other types of cardiac stress and found that deletion of miR-22 rendered mice immune to the hypertrophic effect of the β-adrenergic agonist isoproterenol (17).

A similar role was attributed to the miRNA family miR-212/132 in a study by Ucar et al. The authors showed that miR-212/132 were upregulated in response to hypertrophic stress in cardiomyocytes both in vitro and in vivo (18). Transgenic mice with cardiomyocyte-specific miR-212/132 overexpression spontaneously developed cardiac hypertrophy and died within approximately 100 days with clinical signs of severe HF. A profound cardiac phenotype could also be observed in zebrafish, indicating a highly conserved role for miR-212/132 in the regulation of cardiac function. Moreover, miR-212/132^−/− mice were protected from cardiac hypertrophy in response to left ventricular pressure overload, as opposed to wild-type animals. The anti-hypertrophic transcription factor FoxO3, which is also downregulated in human HF patients, was confirmed as a direct target of miR-212/132. Repressed FoxO3 signaling, via calcineurin and nuclear factor of activated T-cells (NFAT), was proposed as a possible molecular mechanism for the observed effects.

Recent evidence suggest that the deterioration of cardiac function during HF is driven in part by inflammatory processes (19). Heymans et al. showed infiltration of monocytes and macrophages in the hearts of mice subjected to pressure overload (20) and investigated the role of miR-155, one of the most abundant and functionally important miRNAs in inflammatory cells (21), in the pathology of HF. Genetic deletion of miR-155 prevented infiltration of leukocytes into the hearts of mice subjected to pressure overload, and the effect was shown to be mediated by derepression of the anti-inflammatory mediator Socs1 in macrophages. Mutant mice showed preserved cardiomyocyte morphology and contractile function in two different models of HF (Angtioensin II treatment and TAC). In order to translate the findings to human HF, the levels of miR-155 in the hearts of hypertrophic patients with aortic stenosis were assessed and found to be negatively correlated with cardiac function and positively correlated with average wall thickness. In a later study, cardiomyocyte-specific expression of miR-155 was also shown to be involved in pathological processes of HF (22). The authors revealed that miR-155 is involved in the calcineurin-dependent cardiac hypertrophy pathway, possibly through its direct target Jarid2.

miRNAs as therapeutic targets in heart failure

Modulation of miRNA expression for therapeutic purposes has been a concept subject to extensive research in recent years and there are now tools to target specific miRNAs (extensively reviewed by Van Rooij and Olson) (23). Briefly, aberrantly expressed miRNAs can be inhibited by modified antisense oligonucleotides, the so-called “antimiRs” or “antagomiRs,” whereas restoration of pathologically low miRNA levels can be achieved by the introduction of synthetic miRNA mimics. Although oral administration of these oligonucleotides is ineffective because the size and charge of the molecules limit intestinal uptake, succesful delivery via parenteral routes have been shown for most organs, including heart. The molecules are usually taken up by the tissues within hours through an as of yet unknown mechanism. The cellular localization of the oligonucleotides is also unknown, as is the proportion of the molecules that are actually functional.

Thum et al. showed that miR-21 is overexpressed in cardiac fibroblasts of failing hearts and can drive interstital fibrosis and cardiac remodeling in mice. Accordingly, intravenous injection of anti-miR-21 for three days attentuated cardiac fibrosis and hypertrophy in mice subjected to TAC (24). The authors also showed a therapeutic effect of anti-miR-21 injections in mice with established TAC-induced HF.

As mentioned above, miR-208 constitutes an interesting target for treatment of HF. Using Dahl salt-sensitive rats on a high-salt diet as a model of hypertension-induced HF, Montgomery et al. investigated the effect of anti-miR-208 treatment (25). Systemic delivery of anti-miR-208 was initiated one week post diet and rats showed a dose-dependent reduction in miR-208 expression irrespective of route of administration (intravenous, intraperitoneal, or subcutaneous). Treated animals had dramatically reduced mortality compared with control animals, and observations of cardiac structure and function revealed substantial reduction in cardiac remodeling, fibrosis, and isovolumic relaxation time in anti-miR-208-treated rats. Importantly, considering reports of a crucial role for miR-208 in cardiac conductance (26), animals treated with anti-miR-208 did not show any signs of conductance defects. Animals had a high tolerance for the drug regardless of administration route and showed no signs of liver toxicity or other side effects.

Bernardo et al. focused on miR-652 as a therapeutic target in a TAC mouse model of HF. miR-652 was found to be upregulated after induction of myocardial infarction as well as in TAC-induced HF (27). After 4 weeks of TAC, animals with established left ventricular hypertrophy and cardiac dysfunction were either given anti-miR-652 or a scrambled control subcutaneously over a period of 3 days. After 8 weeks, mice given anti-miR-652 had an attenuated hypertrophic response, and improved fractional shortening and left ventricular end-systolic dimension. Moreover, cardiac-stress-response-associated gene expression and cardiomyocyte apoptosis were reduced in anti-miR-652-treated mice. miR-652 was found to be silenced in other organs also, but there were no notable effects on organ weight or morphology.

Serca2 is a key Ca²⁺-transporter in cardiomyocyte excitation–contraction coupling, and HF is associated with reduced Serca2 activity. Beneficial effects of restoring Serca2 expression in failing hearts via gene therapy has been seen in animal models and human patients (28). Based on this knowledge, Wahlquist et al. aimed to identify Serca2-suppressing miRNAs that might be used as therapeutic targets (29). From a whole-genome screen, the authors identified miR-25 as the most potent Serca2 suppressor. miR-25 was also shown to be expressed in cardiomyocytes and induced upon pressure overload in mice. Mice with TAC-induced HF were administered anti-miR-25 and revealed substantial improvement in cardiac function and survival within 2 months. A summary of the miRNAs investigated as therapeutic targets in animal models of HF can be found in Table 1.

Table I. miRNAs as therapeutic targets in animal models of HF.

Target	Model	Treatment	Outcome	Mechanism	References
miR-21	3 Weeks of TAC in mice	Antagomir-21 through jugular vein catheter for 3 days	Attenuation of interstitial fibrosis and cardiac hypertrophy	Derepression of SPRY1 in cardiac fibroblasts	Thum et al. (2008)
			Normalized fractional shortening
miR-208	Dahl rats on a high-salt diet	Tail vein injection of LNA-anti-miR every two weeks	Reduction in isovolumic relaxation time	Reversal of myosin switching	Montgomery et al. (2011)
			Reduction in cardiac hypertrophy
			Reduction in periartieriolar fibrosis
miR-652	4 Weeks of TAC in mice	Subcutaneous administration of LNA-anti-miR for 3 days	Improved fractional shortening/LVESD	Derepression of Jagged1 and Notch signaling	Bernardo et al. (2014)
			Reduced cardiomyocyte apoptosis
			Decreased cardiac fibrosis
miR-25	8 Weeks of TAC in mice	Tail vein injection of anti-miR daily for 3 days, followed by weekly injections for 3 weeks	Restoration of ESPVR and EF	Derepression of SERCA2a and improved Ca2 + uptake	Wahlquist et al. (2014)
			Increased survival
			Reduction in fibrosis
			Normalization of cardiomyocyte size

TAC: Transverse aortic constriction, LNA: Locked nucleic acid, LVSED: Left ventricular end-systolic dimension, ESPVR: End-systolic pressure–volume relationship, EF: Ejection fraction.

miRNA in the diagnosis of heart failure

Correct diagnosis of HF requires extensive investigation of the heart, including echocardiographical, cardiac magnetic resonance imaging or invasive catheterization analyses, but plasma biomarkers such as natriuretic peptides can be helpful in the process, especially in the acute setting (30). Much attention has been given to miRNAs as blood-borne biomarkers for different diseases in recent years. miRNAs have not only been shown to be released into the blood circulation upon tissue injury, stress, or other pathological processes, but are also remarkably stable in biofluids such as plasma and urine, and can be detected with high sensitivity and specificity (31–33).

One of the first studies addressing the usefulness of circulating miRNAs as biomarkers for HF was conducted in 2010 by Tijsen et al. The authors profiled the global miRNA content in plasma from HF patients and healthy controls, and validated their findings in dyspnea patients with and without HF (34). miR-423–5p was significantly elevated in the blood circulation of HF patients and was found to be an accurate predictor of HF both compared with non-HF dyspnea patients and healthy controls. This finding was confirmed in a slightly larger patient sample by Goren et al. miR-423-5p, along with miR-320a, miR-22, and miR-92b were found to be elevated in the blood circulation of HF patients (35). The cumulative level of these four miRNAs could distinguish HF patients from controls with 90% sensitivity and specificity. However, a later study failed to show an association between plasma levels of miR-423-5p and indices of LV function and remodeling during a 1-year period after MI in a relatively large patient population, suggesting that the usefulness of miR-423-5p in the diagnosis of HF might be limited (36).

Corsten et al. found that the cardiac-enriched miRNA miR-499-5p was slightly elevated in the blood circulation of HF patients (37) and was associated with death or development of HF post-MI in a study by our laboratory (38). We could also see a negative correlation between the plasma levels of miR-499 and cardiac function, as measured by left ventricular ejection fraction (LVEF) (38).

Fukushima et al. focused on three miRNAs with different tissue specificities: miR-122 (liver), miR-126 (endothelium), and miR-499 (heart) (39). They found that miR-126 correlated negatively with disease severity, assessed with the New York Heart Asssociation Functional Classification (NYHA). In line with this finding, patients who improved from NYHA IV to NYHA III also showed increased miR-126 levels.

In a recent study, Vogel et al. examined miRNA expression profiles in the whole blood from non-ischemic systolic HF patients (40). Receiver-operating characteristic analyses revealed that miR-558, miR-122*, and miR-520d-5p displayed the highest discriminatory power, with area under curve (AUC) values in the range of 0.7. A miRNA signature of eight miRNAs reached a specificity of 66%, a sensitivity of 74%, and an AUC of 0.81. When comparing single miRNAs with the established HF biomarker NT-proBNP, miR-622, miR-520d, and miR-519e* showed slightly improved performance. A summary of the miRNAs assessed as biomarkers for HF can be found in Table 2.

Table II. miRNAs as biomarkers for HF.

Study	miRNA(s)	Study population	Diagnostic accuracy (AUC from ROC analysis)
Tijsen et al. (2010)	miR-423-5p	HF cases, n = 30	HF vs healthy controls 0.91
		Non-HF dyspnea cases, n = 20	HF vs non-HF dyspnea 0.83
		Healthy controls, n = 39
Goren et al. (2012)	miR-423-5p	Chronic HF cases, n = 30	Combined miRNA-score, HF vs controls: 0.9
	miR-320a	Healthy controls, n = 30
	miR-22
	miR-92b
Corsten et al. (2010)	miR-499	Acute HF, n = 33	miR-499-5p 2-fold elevated in AHF, (p < 0.05)
		Healthy controls, n = 20	ROC N/A
Vogel et al. (2013)	miR-520d-5p	Non-ischemic HF cases, n = 53	Combined miRNA signature, HF vs controls: 0.81
	miR-558	Healthy controls, n = 39
	miR-122*
	miR-200b*
	miR-622
	miR-519e*
	miR-1231
	miR-1228*

HF: Heart failure, AUC: Area under the curve, ROC: Receiver-operating characteristic, N/A: not available.

Conclusion

Two decades have passed since the discovery of miRNA, and our understanding of its roles in human health and disease is still accelerating. Recently, there has been a realization that while miRNAs usually exert moderate effects under homeostatic conditions, it is in situations of stress and injury that their function becomes pronounced (41). It is therefore no surprise that miRNAs play crucial roles in the cardiac stress response, orchestrating a shift in gene expression programs to meet the increased demands on cardiac output. It is evident from the literature reviewed here that miRNAs are implied in the processes leading to heart failure, including cardiac remodeling, fibrosis, and inflammation (Figure 1). Inhibiting some of these miRNAs with synthetic antimiRs have had beneficial effects on cardiac function and survival in animal models of heart failure. Although this strategy holds a lot of promise, considerable efforts are needed in order for this kind of therapy to reach the clinic. Many question marks remain, not in the least over potential off-target effects, considering that a single miRNA can have several hundred putative mRNA targets. Additionally, there are issues regarding specificity, modes of delivery, and potential toxicity that need thorough investigation. Nevertheless, seeing that the first miRNA-targeting drug has now successfully completed a phase-2a clinical trial (42), the future for miRNAs as therapeutic targets is exciting.

Figure 1. Overview of miRNAs involved in the pathogenesis of heart failure. miRNAs and their respective target mRNAs in different cell types are shown.

Declaration of interest: The authors report no declarations of interest. The authors alone are responsible for the content and writing of the paper.