Cell type–specific microRNA therapies for myocardial infarction

Abstract

Current interventions fail to recover injured myocardium after infarction and prompt the need for development of cardioprotective strategies. Of increasing interest is the therapeutic use of microRNAs to control gene expression through specific targeting of mRNAs. In this Review, we discuss current microRNA-based therapeutic strategies, describing the outcomes and limitations of key microRNAs with a focus on target cell types and molecular pathways. Last, we offer a perspective on the outlook of microRNA therapies for myocardial infarction, highlighting the outstanding challenges and emerging strategies.

INTRODUCTION

Heart disease remains as the leading cause of mortality worldwide, with myocardial infarction (MI) affecting more than 700,000 individuals annually in the United States alone (1). Because the adult heart lacks ability to innately repair and regenerate after injury, MI results in permanent loss of myocardial cells (2). The damaged heart undergoes pathological remodeling, leading to reduced contractile function and often heart failure (3). Standard therapies mainly focus on revascularizating the occluded artery to salvage as much of the myocardium as possible but fail to adequately recover injured myocardium. Any advances in the treatment of MI would thus have major clinical significance.

The human heart is composed of an array of different cell types including cardiomyocytes (CMs), cardiac fibroblasts (CFs), endothelial cells (ECs), and immune cells. Each cell type contributes to cardiac function in a distinct way, with intercellular communication and coordination being vital to maintaining normal organ function (Fig. 1). The pathological changes in cardiac remodeling after MI involve each of these cell types and multiple molecular mechanisms during four partially overlapping phases: the inflammatory, proliferative, maturation, and remodeling phases (3). The inflammatory phase is initiated by massive cell death in the infarct area. Although CFs degrade the extracellular matrix (ECM), changes in EC-mediated vascular permeability allow immune cells to migrate into the injured area and facilitate clearance of damaged cells. In the proliferative phase, inflammation is dampened by macrophage phenotypic switching, and reparative processes begin. Fibroblasts and ECs proliferate, deposit collagen, and establish a new microvascular network in the infarct area. In the maturation phase, ECM proteins secreted by activated fibroblasts undergo cross-linking to form a stable scar. Last, during long-term remodeling, maladaptive processes including cardiomyocyte pathological hypertrophy, remote fibrosis, and capillary rarefaction occur to contend with the stresses arising from the failing heart (Fig. 1).

Fig. 1. Summary of myocardial infarct progression with a focus on the role of each cell type.

Immediately after infarction, the inflammatory phase begins and is characterized by irreversible cell death of all cell types in the infarct and the initiation of the inflammatory process through cytokine release. Next, the proliferative phase is characterized by the activation of CFs and the initiation of scar development. Macrophages also undergo polarization and switch to an anti-inflammatory phenotype during this phase. The maturation phase is characterized by cardiomyocyte hypertrophy in response to increased cardiac demand and the maturation of the scar that forms over the infarcted area. Last, long-term remodeling is characterized by pathological changes including maladaptive cardiomyocyte hypertrophy, chronic overactivation of CFs, decreased vascular density, and reemergence of an inflammatory response.

Recently, microRNAs (miRNAs) have emerged as a promising therapy for MI. miRNAs are small noncoding RNAs that regulate gene expression at the posttranscriptional level (4). Mature, single-stranded miRNAs are incorporated into miRNA-induced silencing complexes, the functional unit that targets mRNAs with near-perfect base pairing to inhibit gene expression (5). Known to interact with the majority of mammalian protein-coding genes, miRNAs are powerful mediators of a diverse spectrum of processes, both during normal heart development and physiology and during cardiovascular disease progression (6). miRNAs are implicated in every phase of MI progression; in response to ischemia, miRNAs in both mouse and human hearts have been shown to be dysregulated, contributing to the progression of many pathological processes.

For treatment of MI, miRNAs represent particularly attractive therapeutic targets due to several unique characteristics. Most important is the pleotropic ability of a single miRNA to regulate multiple pathologically disrupted biological pathways across different cell types (7). This is in stark contrast to traditional drug-based approaches that target singular molecules and pathways. Such a pleiotropic approach is especially powerful for treatment of MI because injury is not instigated by a single genetic link or biological process but rather by multiple coordinated processes. In addition, miRNAs are ideal therapeutic targets because they are small, precisely defined nucleic sequences for which mimics or antisense oligonucleotides (ASOs) can effectively and efficiently be designed with high affinity and specificity. Last, when compared to cell-based approaches, miRNA therapies may provide similar benefits without the challenges of immune rejection and poor cell engraftment.

Although miRNA therapies have potential for treating MI, a major challenge that limits their clinical advancement is a lack of understanding of the molecular mechanisms behind their therapeutic properties. Specifically, the pleotropic effects of a miRNA on each cell type and on different biological pathways within the heart must be carefully delineated to facilitate clinical translation. In this Review, we examine recently published studies describing the therapeutic manipulation of miRNAs in the treatment of MI. We describe the effect of miRNAs on each cell type and how miRNAs directly target molecular pathways to modulate cell type–specific behavior. Last, we discuss key prospects and challenges for developing miRNA therapies.

CHARACTERISTICS OF miRNA THERAPIES

Focusing on literature within the past 3 years and foundational studies from the past 10 years, we identified a total of 213 relevant studies detailing 116 unique miRNAs (Fig. 2). Therapeutic miRNA targets were identified by two methods: (i) miRNA profiling after MI, which revealed substantial dysregulation of miRNAs natively expressed in the heart; and (ii) miRNA screening in in vitro injury models, which identified potential proregenerative miRNAs not natively expressed in the heart. To understand cell type–specific effects of miRNA therapies, we organized the reviewed miRNAs by the cell types they target. For each study, we extracted a set of characteristics describing each miRNA. miRNAs were defined as beneficial if treatment with miRNA mimics resulted in improved recovery after MI and as harmful if treatment with miRNA inhibitors resulted in improved recovery after MI. We noted the model systems each study used in vitro and in vivo as well as the delivery of the miRNA therapies (miRNA formulation and the method of administration). For each miRNA studied, we detail the target pathway and the specific target molecule.

Fig. 2. Summary of miRNAs surveyed in this Review.

For each study, in this Review, we categorized the overall effect of the miRNA studied, whether the miRNA was shown to affect multiple cell types (pleiotropy), which cell types were studied, and the physiologic process that the miRNA was shown to target. The numbers following each division represents the number of studies that is in that category.

Using these characteristics, we assigned a clinical applicability score representing a qualitative measure of the evidence supporting the therapeutic application of each miRNA reviewed. This score is denoted by the footnotes in Table 1. “*” indicates therapeutic efficacy in in vitro cell-based models; “**” indicates therapeutic efficacy in in vivo models with target pathway identification and characterization; and “***” indicates therapeutic efficacy in multiple model systems and well-defined mechanisms of action including comprehensively validated direct targets. Table 1 presents a select list of key miRNAs for each cell type. The complete list of surveyed miRNAs is presented in table S1 and summarized in table S2.

Table 1. Select cell type–specific miRNA therapies in the treatment of MI.

Complete list of miRNAs reviewed in table S1. IRI, ischemia-reperfusion injury; AngII, angiotensin II; TAC, transverse aortic constriction; IV, intravenous; IM, intramyocardial; IP, intraperitoneal; I. Ventr, intraventricular; EV, extracellular vesicle; Transf, transfection; Lenti, lentiviral; EndMT, endothelial-to-mesenchymal transition; Mito, mitochondrial; ROS, reactive oxygen species.

miRNA	Effect	Injury model	Delivery method	Target process	Target pathway	Target molecule	Clinical application	Reference
Cardiomyocytes
miR-1	Harmful	In vivo IRI	IV	Apoptosis	PKC apoptosis	PKCε	**	(43)
miR-9	Harmful	In vivo ischemia	IM	Proliferation	FSTL	FSTL	***	(73)
miR-22	Harmful	In vivo ischemia	IP	Autophagy	Autophagy	PPARα	***	(59)
miR-302/367	Beneficial	In vivo ischemia	IV	Proliferation	Hippo-YAP	MST1	***	(69)
miR-762	Harmful	In vivo IRI	IV	Apoptosis	ROS production	ND2	***	(49)
miR-873	Beneficial	In vivo IRI	IV	Necrosis	Necroptosis	RIPK1/RIPK3	***	(53)
Let-7a	Beneficial	In vivo AngII	I. Ventr.	Hypertrophy	Undefined	CaM	***	(80)
Cardiac fibroblasts
miR-21	Harmful	In vivo ischemia	IM	Fibrosis	ERK/MAPK	SPRY1	***	(112)
miR-21	Harmful	In vivo TAC	IV	Apoptosis	ERK/MAPK	SPRY1	***	(111)
miR-92a	Harmful	In vitro TGF-β	EV	Fibrosis	TGF-β	SMAD7	*	(106)
miR-101	Beneficial	In vivo ischemia	I. Ventr.	Fibrosis	TGF-β	cFOS	***	(96)
miR-155	Beneficial	In vivo genetic	NA	Proliferation	RAS	SOS1	**	(178)
Endothelial cells
miR-32	Harmful	In vitro	Transf.	Proliferation	Undefined	KLF2	*	(228)
miR-155	Harmful	In vivo ischemia	IM	Angiogenesis	AMPK	RAC1	***	(142)
miR-324	Beneficial	In vitro H2O2	Transf.	Apoptosis	Mitofission	MtFR1 (indirect)	*	(135)
miR-532	Beneficial	In vivo ischemia	IM	Angiogenesis	EndMT	PRSS23	***	(151)
Inflammatory cells
miR-24	Beneficial	In vivo ischemia	EV	Inflammatory infiltration	Intrinsic pathway	BIM	***	(161)
miR-155	Harmful	In vivo ischemia	IV	Inflammatory activation	NF-κB	SOCS1	***	(175)
miR-182	Beneficial	In vivo IRI	EV	Inflammatory resolution	PI3k/AKT/mTOR	TLR4	**	(162)
miR-224	Beneficial	In vivo ischemia	Lenti.	Inflammatory activation	TGF-β	Undefined	*	(157)

^*Indicates therapeutic efficacy in in vitro cell-based models.

^**Indicates therapeutic efficacy in in vitro and in vivo models with target pathway identification and characterization.

^***Indicates therapeutic efficacy in multiple model systems and well-defined mechanisms of action including comprehensively validated direct targets.

Most miRNAs play both beneficial and harmful roles in the progression of MI. In addition, a majority of miRNAs affect multiple cell types, thus demonstrating the important pleiotropic effects of miRNAs when used in cardiac therapy. Here, we first highlight post-MI events by cell type, briefly describing how key processes in each cell type affect their phenotype and function. After providing this contextual framework, we examine the recent advances in miRNA therapies targeting each cell type.

ROLE OF miRNAs IN CARDIOMYOCYTE RESPONSE TO MYOCARDIAL INFARCTION

Cardiomyocytes are the functional cells of the heart, responsible for generating contractile force. Because of their large energy demands, cardiomyocytes are highly susceptible to death upon loss of blood supply. Cardiomyocytes have limited proliferative ability, and those lost during MI cannot be replaced (8). Thus, two major approaches to therapeutically target cardiomyocytes after MI are prevention of cardiomyocyte cell death after ischemic stress and induction of cardiomyocyte proliferation after resolution of injury.

Cardiomyocyte response to myocardial infarction

Apoptosis and necrosis

The primary pathways of regulated cardiomyocyte cell death during MI are apoptosis and necrosis, both of which have distinct molecular mechanisms and cellular conditions. Mechanistically, apoptosis and necrosis can both be subdivided into mitochondria- and receptor-mediated pathways, known as the intrinsic and extrinsic pathways during apoptosis and as mitochondrial necrosis and necroptosis during necrosis. During MI, the intrinsic pathway is initiated when free radical generation and low adenosine triphosphate (ATP) lead to mitochondrial stress and subsequent perturbation of the B cell lymphoma 2 (BCL-2) family of proteins. Simultaneously, mitochondrial necrosis is initiated when intracellular Ca²⁺ overload alters the dynamics of the mitochondrial membrane, leading to a disruption in permeability. This results in a rapid decrease in ATP, which prevents the cell from carrying out necessary repair functions, leading to organelle dysfunction and plasma membrane rupture. The extrinsic pathway of apoptosis and necroptosis are both triggered during MI by release of stimuli outside the cell that activates death receptors on the cell surface such as Fas ligand activating Fas receptor (FasR) and tumor necrosis factor–α (TNFα) activating TNF receptor 1 (TNFR1) (9).

Autophagy

In response to the ischemic or metabolic stress during MI, cardiomyocytes can activate the process of autophagy, which allows them to catabolize damaged or dysfunctional macromolecular structures by lysosomal degradation to maintain cellular homeostasis. Primary players in cardiomyocyte autophagy include the mammalian target of rapamycin (mTOR) and AMP-activated protein kinase (AMPK) pathways (10). Although undoubtedly critical in cardiomyocyte response to MI, the role of autophagy in mediating the extent of myocardial injury is uncertain.

Proliferation and remodeling

Historically, adult cardiomyocytes have been described as postmitotic, but this paradigm has shifted, with studies indicating that cardiomyocytes proliferate to a small extent in the adult myocardium (11). Although limited, exogenously promoted cardiomyocyte proliferation has become a potential therapeutic strategy for heart failure. Over the past decade, several promising targets have emerged that induce cardiomyocyte proliferation, including Hippo–Yes-associated protein (YAP), homeodomain-only protein (HOPX), and follistatin-related protein 1 (FSTL) (12–14).

Because of the loss of functional cells after MI, cardiomyocytes undergo hypertrophy and morphological changes to adapt to the physical demands of the injured heart. Although initially compensatory, prolonged increases in functional demands eventually lead to pathological remodeling and a decrease in cardiomyocyte function. Infarcted tissue also interrupts gap junctions between cardiomyocytes, causing aberrant electrical conduction leading to arrhythmias (15).

Recent advances in miRNA therapies targeting cardiomyocytes

Most studies (132 of 213 studies) describing the therapeutic application of miRNAs to MI have cardiomyocytes as the primary target cell type. These studies have focused on identifying and using miRNAs capable of either preventing cardiomyocyte death (88 of 132 studies) or promoting cardiomyocyte proliferation (18 of 132 studies). Here, we summarize key therapeutic miRNAs and their mechanisms of action in cardiomyocytes (Fig. 3).

Fig. 3. Summary of miRNAs that target cardiomyocytes after MI.

miRNAs have been shown to regulate apoptosis, necrosis, and autophagy-mediated cardiomyocyte cell death. For cardiomyocyte proliferation, miRNAs have been shown to modulate multiple proliferative pathways as well as the cell cycle directly. Last, miRNAs have also been shown to regulate important processes including hypertrophy, arrhythmia, and inflammation. Select well-characterized miRNAs, their targets (written above the arrows), and the processes that they regulate are shown (i.e., miR-19b prevents apoptosis through BIM signaling). miRNAs prefaced with “anti-” denote that the inhibition of the miRNA is therapeutic (i.e., the inhibition of miR-1 prevents apoptosis through BCL-2 signaling). For the complete list of miRNAs reviewed that affect cardiomyocytes, please see table S1.

Apoptosis

Many miRNAs have been identified that regulate components of apoptosis pathways, with beneficial miRNAs inhibiting these pathways to prevent cardiomyocyte apoptosis and harmful miRNAs inducing apoptosis after injury. In early apoptosis, mitochondrial fission leads to activation of mitochondria-mediated cardiomyocyte death. In a series of studies, Wang and colleagues (16) described five miRNAs that target components of mitochondrial fission machinery. They identified two beneficial miRNAs, miR-499 and miR-652, which prevent mitochondrial fission. miR-499 directly inhibits both the α and β isoforms of the calcineurin catalytic subunit, thereby decreasing the accumulation of dynamin-related protein 1 (DRP1), a guanosine triphosphatase required for mitochondrial fission (16). miR-652 inhibits the proapoptotic mitochondrial membrane protein MTP18 (17). In contrast, miR-361 and miR-539 were identified as harmful miRNAs that inhibit prohibitins 1 and 2, respectively, increasing mitochondrial fission and apoptosis after MI (18, 19). miR-421 was also identified as harmful after MI as it directly targets the mitochondrial Ser/Thr kinase PTEN induced kinase 1 (PINK1), which inhibits mitochondrial fission (20).

Harmful miRNAs including miR-195 and miR-1 have been shown to directly inhibit the antiapoptotic protein BCL-2 that regulates the intrinsic pathway (21, 22). Alternatively, beneficial miRNAs such as miR-19b and miR-24 have been shown to directly inhibit the proapoptotic protein Bcl-2-like protein 11 (BIM) (23, 24). In a foundational study, Qian et al. (24) demonstrated that the overexpression of miR-24 via intramyocardial injection of a miR-24 mimic attenuated ischemia-induced injury and restored cardiac function by directly reducing BIM expression. miRNAs affect other components of the intrinsic pathway: For example, miR-125b inhibits proapoptotic Bcl-2 homologous antagonist/killer 1 (BAK1), miR-17 inhibits proapoptotic apoptotic protease activating factor-1 (APAF-1), and miR-27a inhibits Bcl-2 interaction protein 3 (BNIP3) (25–27). To modulate the extrinsic pathway, miR-133b can down-regulate the death receptor FasR, and the effector caspase-3 can be inhibited by miR-1192 (28, 29). Last, miR-327 has been shown to inhibit apoptosis repressor with caspase recruitment domain (ARC), a potent repressor of the intrinsic and extrinsic signaling pathways (30).

In cardiomyocytes, multiple pathways respond to ischemic stress and injury after MI and ultimately determine whether the cell undergoes apoptosis. The PI3k/AKT pathway plays a major role in controlling cell survival and the inhibition of programmed cell death through stimulatory phosphorylation of prosurvival genes and the inhibitory phosphorylation of proapoptotic genes (31). Many miRNAs modulate the PI3k/AKT pathway in cardiomyocytes after MI. Specifically, the central AKT inhibitor phosphatase and tensin homolog (PTEN) can be directly targeted by miRNAs miR-146b, miR-182, miR-26a, miR-494, and miR-93 (32–35). Song et al. (36) demonstrated that miR-320 directly inhibits the ligand insulin growth factor 1 (IGF-1) after ischemia-reperfusion injury (IRI), thus preventing IGF receptor-mediated activation of the PI3k/AKT pathway. Restoring IGF-1 function using a lentivirus expressing miR-320 inhibitor led to a decrease in the number of apoptotic cardiomyocytes and preserved cardiac function (36). Another notable pathway regulating apoptosis centers around the molecule programmed cell death 4 (PDCD4), which is up-regulated during apoptosis and functions as a proapoptotic inhibitor of gene transcription and translation. During MI, many miRNAs directly inhibit PDCD4 and prevent apoptosis in cardiomyocytes including miR-200, miR-499, miR-532, and, importantly, miR-21, a miRNA with notable pleiotropic effects (37–42).

Additional pathways implicated in cardiomyocyte apoptosis include protein kinase C (PKC), nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), Notch, and reactive oxygen species (ROS) homeostasis. In PKC-mediated apoptosis, the harmful miRNAs miR-1, miR-143, and miR-31 have been shown to directly target antiapoptotic protein kinase C epsilon (PRKCE), thereby leading to increased cardiomyocyte apoptosis (43–45). In NF-κB signaling, miR-145a exerts beneficial effects by inhibiting early growth response 1 (EGR1)–mediated NF-κB activation. In Notch signaling, the harmful miRNAs miR-363 and miR-429 directly inhibit the Notch 1 receptor, preventing Notch-mediated antiapoptotic signaling (46, 47). In contrast, miR-322 targets the Notch inhibitor FBXW7, up-regulating Notch signaling and preventing apoptosis after MI (48). In the regulation of ROS, miRNAs can directly affect mitochondrial respiration. For example, the harmful miR-762 inhibits the mitochondrial nicotinamide adenine dinucleotide + hydrogen (NADH) dehydrogenase subunit ND2 leading to increased ROS production, whereas the beneficial miR-183 inhibits the mitochondrial membrane ion channels voltage-dependent anion channel 1 (VDAC1) leading to decreased ROS production (49, 50). In addition, Su et al. (51) showed that miR-132 inhibits histone deacetylase HDAC3, preventing its ability to up-regulate genes that promote ROS accumulation in cardiomyocytes.

Necrosis

Large numbers of cardiomyocytes undergo necrotic cell death after ischemic injury, providing the main stimulus for postinfarction inflammatory activation. Here, beneficial miRNAs prevent cardiomyocyte necrosis, whereas harmful miRNAs induce necrosis after injury. The central component of necroptosis activation is the assembly of the receptor-interacting protein kinase 1 (RIPK1)–RIPK3-mixed lineage kinase-like (MLKL) signaling complex. Two miRNAs, miR-105 and miR-873, prevent necroptosis through the direct down-regulation of RIPK proteins (52, 53). In addition, miRNAs can modulate the expression of death receptors responsible for initiating necroptosis. The beneficial miR-223 inhibits the receptors TNFR1 and DR6, thereby preventing necroptosis; whereas the harmful miR-103/107 inhibits the antinecrotic receptor Fas-associated death domain protein, thereby activating necroptosis (54, 55). Last, the beneficial miR-874 inhibits caspase-8 preventing its ability to activate the RIPK signaling complex (56).

Autophagy

Studies have identified miRNAs that modulate post-ischemic autophagy, sometimes with contradictory effects on cardiac function. For example, Wang et al. (57) demonstrated that miR-188 suppresses autophagy-mediated apoptosis after MI through direct inhibition of the autophagy-related gene ATG7. Similarly, Liu et al. (58) showed that miR-93 protects cardiomyocytes from apoptosis through inhibition of ATG7. In a comprehensive study, Gupta et al. (59) identified miR-22 as a potent inhibitor of cardiac autophagy through its targeting of peroxisome proliferator–activated receptor α (PPARα), a nuclear receptor known to activate autophagy. However, inhibition of miR-22 leading to the activation of autophagy after MI resulted in improved postinfarction remodeling and improved cardiac function (59). Similarly, Ucar et al. (60) demonstrated that the miR-212/132 family inhibited autophagy and that their inhibition restored the autophagic response, resulting in the rescue of heart failure in mice. Other studies have identified miRNAs that can either promote or inhibit autophagy including miR-142, miR-145, miR-221, miR-301, miR-18a, miR-199a, miR-30a, and miR-558; however, the ultimate therapeutic effects of these miRNAs seem to depend on their individual targets and not the overall activation or inactivation of autophagy (61–68).

Proliferation

Cardiac function can be restored after MI by generation of new cardiomyocytes. To this end, multiple miRNAs have been identified that can regulate cardiomyocyte proliferation. Here, beneficial miRNAs promote, whereas harmful miRNAs prevent cardiomyocyte proliferation after injury.

The Hippo-YAP pathway is a critical regulator of organ size and growth that plays an important role in cardiomyocyte proliferation (12). In a foundational study that delivered a transient proliferative stimulus using miRNA mimics, Tian et al. (69) showed that the delivery of miR-302/367 to the infarcted myocardium promoted cardiomyocyte proliferation and regeneration by targeting macrophage stimulating 1 (MST1), a central Hippo-YAP player. HOPX is a key regulator of heart development and induces cardiomyocyte proliferation when overexpressed (14). In a series of studies, the Giacca lab determined that miR-590 and miR-199a both promote cell cycle reentry in adult cardiomyocytes by targeting HOPX (70). They subsequently demonstrated that the proproliferative effect of miR-199a is also dependent on its regulation of the Hippo-Yap pathway (71). Overexpression of miR-199a using adeno-associated virus 6 in a swine model of MI resulted in unimpeded cardiomyocyte proliferation that progressively reduced the cardiac scar size but eventually led to sudden cardiac death in 70% of treated animals (72). The loss of FSTL1 has been shown to be a maladaptive response to injury, whereas its restoration resulted in increased numbers of proliferating cardiomyocytes (13). Recently, Xiao et al. (73) has demonstrated that miR-9 directly inhibits FSTL1 after MI and that the inhibition of miR-9 results in restored FSTL1, increases cardiomyocyte proliferation, and preserves cardiac function.

In parallel to targeting these canonical pathways, a set of miRNAs prevents proliferation by directly targeting cell cycle regulators such as cyclins. miR-34a and let-7i inhibit cyclin D1 and cyclin D2, respectively, suppressing their functions through the cell cycle (74, 75). In addition, miR-294 inhibits checkpoint kinase Wee1, preventing its suppression of the CDK1/cyclin B1 complex and cell cycle reentry (76). Last, miR-128 inhibits the chromatin modifier SUZ12, preventing it from activating positive cell cycle regulators cyclin E and CDK2 (77).

Other effects

Beyond cardiomyocyte cell death and proliferation, miRNAs can affect other facets of cardiac physiology. The harmful miRNAs miR-1231 and miR-223 induce arrhythmias in the heart after MI through inhibition of the ion channels CACNA2D2 and KCND2, respectively (78, 79). The beneficial miRNAs Let-7a and miR-206 prevent pathological cardiomyocyte hypertrophy by targeting calmodulin (CAM) and forkhead box protein P1 (FOXP1), respectively (80, 81). Meanwhile, the harmful miRNAs of the miR-212/132 family can induce hypertrophy through inhibition of FOXO3 and subsequent hyperactivation of NFAT signaling (60). A series of miRNAs have been shown to modulate the ability of cardiomyocytes to activate the inflammatory cascade: miR-128 targets SOX7, resulting in increased inflammation; miR-135b targets caspase-1, resulting in reduced inflammation; and miR-145 targets CD40, also resulting in reduced inflammation (82–84).

ROLE OF miRNAs IN FIBROBLAST RESPONSE TO MYOCARDIAL INFARCTION

After MI, CFs become activated and differentiated into myofibroblasts, which play an integral role in the rapid formation of a scar necessary to prevent ventricular wall rupture. However, excessive CF activation, proliferation, and ECM deposition after MI contribute to pathological cardiac fibrosis, which can exacerbate injury and lead to heart failure. Reversion of this activated phenotype with miRNAs is a key approach toward attenuating this pathological response.

Fibroblast responses to myocardial infarction

Inflammation

Immediately after MI, fibroblasts produce proinflammatory cytokines in response to cardiomyocyte death and inflammatory milieu. Interleukin-1α (IL-1α) stimulates CFs to secrete TNFα and IL-1β (85). Fibroblast signaling acts as a source of IL-1β positive feedback, attracting immune cells into the infarct zone (86). Cytokine-activated inflammatory fibroblasts modulate the secretion of proteases including matrix metalloproteinases (MMPs) that are essential for clearing the infarct of damaged matrix debris (87).

Proliferation

The process that has drawn the most focus in fibroblast repair of the heart is their differentiation into myofibroblasts and migration into the infarct zone. Transforming growth factor–β (TGF-β) is a key regulator of this phase (88). After MI, TGF-β1 is generated by macrophages producing angiotensin II (AngII). AngII has autocrine function, causing the up-regulation of TGF-β1 (89). TGF-β1 signaling involves a complex cascade of proteins including activating and inhibitory SMADs, with pleiotropic effects. TGF-β suppresses MMPs (90) while significantly increasing the production of collagens type 1 and 3, causing ECM synthesis (91). Activated myofibroblasts express α-smooth muscle actin (αSMA), allowing for scar contraction. Notably, TGF-β1 is secreted at an increased rate globally in the myocardium after injury, not just directly at the site of infarction (92), indicating widespread ramifications of MI.

Scar maturation

During the maturation phase, ECM proteins secreted by activated fibroblasts are cross-linked to form a stable scar. Over time, the myofibroblast population in the scar decreases (93) through apoptosis and transition into a recently described fibroblast phenotype termed the matrifibrocyte that is capable of maintaining the scar integrity (94). There is increasing evidence that the myofibroblast phenotype is reversible in vitro: CFs isolated from patients with heart failure revert to quiescence upon TGF-β1 inhibition (95). Therefore, the failure of myofibroblasts to deactivate has broad implications for clinical worsening of MI. Overactivation is driven by multifactorial processes, including the persistent secretion of AngII in parts of the heart remote from the site of injury (96). MiRNA therapies targeting fibroblasts seek to break the positive feedback loop of myofibroblast activation, in which the stiffening of the heart matrix induces even more TGF-β secretion.

Recent advances in miRNA therapies targeting cardiac fibroblasts

The inhibition of myofibroblast activity is a major focus of ongoing research with the identification of beneficial miRNAs that prevent and harmful miRNAs that induce fibroblast activation. Particular attention has been put on the TGF-β signaling pathway (15 of 26 studies). Here, we summarize key miRNAs and their mechanisms of action in CFs (Fig. 4A).

Fig. 4. Summary of miRNAs that target fibroblasts, endothelial cells, and immune cells.

(A) In CFs, miRNAs can regulate TGFβ signaling as well as ECM synthesis and pathological proliferation. (B) In ECs, a large number of miRNAs have been shown to regulate angiogenesis, whereas others have been shown to regulate EC apoptosis. (C) In immune cells, miRNAs can regulate multiple aspects of macrophage function including immune cell infiltration, cytokine production, macrophage polarization, efferocytosis, and phagocytosis, as well as angiogenesis signal release. Select well-characterized miRNAs, their targets (written above the arrows), and the processes that they regulate are shown (i.e., miR-24 prevents TGFβ activation through FURIN signaling). miRNAs prefaced with anti–denote that the inhibition of the miRNA is therapeutic (i.e., the inhibition of miR-433 prevents TGFβ activation through AZIN1 signaling). For the complete list of miRNAs reviewed, please see table S1.

TGF-β signaling

Many miRNAs regulate parts of the TGF-β pathway, including its production, cleavage, and signal transduction. The classical upstream regulator of TGF-β, c-Fos, is regulated by miR-101a, with overexpression of miR-101a leading to reduced fibrosis and improved function after MI (96). Concurrently, Zhao et al. (97) identified TGF-β receptor 1 (TGFBR1) as another target of miR-101a, using a rat model to show decreases in TGF-β signal transduction. Another potential target for TGF-β modulation is AZIN1, as the knockdown of AZIN1 up-regulated the TGF-β expression, with miR-433 directly reducing AZIN1 in vivo (98). Transfection of miR-433 into CFs increased fibroblast proliferation and αSMA expression, whereas injection of a miR-433 antagomir in rats increased AZIN1 in post-MI hearts, leading to a reduction in fibrosis. Proteases cleaving latent TGF-β to its active form can also be targeted by miRNAs to reduce fibrosis. Wang et al. (99) showed that miR-24 inhibits furin, one such protease. In vitro overexpression of miR-24 increased TGF-β secretion and Smad2/3 phosphorylation, and miR-24 was found to be underexpressed in MI tissue.

Many miRNAs target the cellular receptors that bind TGF-β. Hong et al. (100) identified TGFBR1 (one of two main receptors for TGF-β) as a target of miR-22. Overexpression of miR-22 reduced AngII activation of CFs in vitro. Liang et al. (101) described a reciprocal loop by which TGF-β up-regulates miR-21, which then inhibits TGFBR3, a negative regulator of TGF-β signaling. The inhibition of TGFBRIII increases TGF-β secretion and Smad3 phosphorylation and therefore collagen secretion. Similarly, Du et al. (102) showed that miR-328 targets TGFBRIII and that an injection of anti–mir-328 can improve cardiac fibrosis post MI in mice.

After binding to the receptor, TGF-β signaling is mediated by decapentaplegic (DPP) homologs (SMAD) family. SMAD2/3 couple to the receptor and are phosphorylated by TGF-β binding, subsequently binding to SMAD4. SMAD6 and SMAD7 inhibit TGF-β signaling. miRNA mimics, which target SMAD2/3/4 and anti-miRs against those that target SMAD6/7, can prevent fibrosis. Yuan et al. (103) showed the suppression SMAD7 by miR-21, with maladaptive response in fibroblasts. miR-34a is up-regulated after MI, and in vivo inhibition of miR-34a reduced post-MI cardiac fibrosis: SMAD4 was identified as a direct target of miR-34a (104). Ischemic exosomes isolated from mouse cardiomyocytes after MI were abundant in miR-92a, which directly targets SMAD7. Overexpression of miR-92a contributed to the activation of fibroblasts (105). These post-MI cardiomyocyte exosomes contain miR-195, which also targets SMAD7 (106).

Last, the downstream effectors of TGF-β, such as connective tissue growth factor (CTGF) (107, 108), can also be targeted by miRNAs. Duisters et al. (109) initially identified both miR-133 and miR-30 as inhibitors of CTGF. Later, Chen et al. (110) delivered miR-30a using adenovirus after MI and observed increased heart function and decreased collagen deposition.

Other pathways

The harmful role of miR-21 in cardiac fibrosis was established by the Thum lab (111, 112), which showed, in a mouse model of heart failure, that miR-21 was dysregulated in fibroblasts but not cardiomyocytes. Inhibition of miR-21 increased the percentage of apoptotic fibroblasts in the failing heart. They demonstrated that miR-21 targets Sprouty1, resulting in increased prosurvival extracellular signal–regulated kinase (ERK)/mitogen-activated protein kinase (MAPK) signaling. Delivery of anti–miR-21 in vivo attenuated fibrosis. Similarly, Cardin et al. (113) demonstrated that miR-21 is up-regulated in atria after MI and contributes to the development of fibrosis and arrhythmia in atrial fibrillation. miR-21 knockdown reduced atrial fibrosis and stabilized electrical conduction.

Yuan et al. (114) identified miR-144 as a PTEN inhibitor in miniature swine. Overexpression of miR-144 or transfection of a PTEN-targeting small-interfering RNA (siRNA) in primary human CFs in vitro increased collagen 1 and αSMA mRNA expression.

Since smooth muscle gene expression is a hallmark of differentiated myofibroblasts, miRNAs that regulate smooth muscle phenotype are potentially therapeutic. miR-143 and miR-145 are transcribed from the same cluster and play a critical role in smooth muscle differentiation (115). Li and colleagues (116) showed that miR-143 is up-regulated in human MI tissue samples, and miR-143 inhibitors reversed the effects of TGF-β stimulation on fibroblasts in vitro. miR-143 directly binds to the 3′ untranslated region of Sprouty3, activating p38, ERK, and c-Jun N-terminal kinase pathways. In another study, Wang et al. (117) showed that miR-145 is sufficient to increase the myofibroblast phenotype of CFs by targeting KLF5, which is a negative regulator of the myocardin serum response factor (SRF) pathway. miR-145 inhibition in vivo reduced αSMA expression but increased scar size, perhaps indicating a decrease in contractile function of the fibroblasts.

The miR-133a family, mir-133a-1, miR-133a-2, and miR-133b, are also key regulators of the SRF pathway. Of these, the two miR-133a genes are identical and specifically expressed in cardiac myocytes, whereas miR-133b is expressed in skeletal muscle (118). miR-133a is down-regulated in the hearts of patients with MI as well as in animal models (119). miR-133a double knockout mice display severe fibrosis and early mortality (118). Expressed specifically in cardiomyocytes, miR-133a appears to act in a paracrine manner, reducing the secretion of profibrotic cytokines. Duisters et al. (109) showed that miR-133a regulates the production of CTGF in cardiomyocytes, a downstream effector of TGF-β signaling.

ECM proteins can also be targeted as the end product of myofibroblast activation. The Olson lab demonstrated that miR-29 is decreased after MI and that it targets ECM protein mRNAs, including collagens, fibrillin, and elastins (120). Inhibition of miR-29 in vivo induced collagen mRNA expression in the heart and other organs. miR-29 was found to correlate to collagen expression after MI (121), and collagen 1 was identified as a target of miR-133a (122) in AngII-induced injury. However, an opposite observation was found with genetic knockout of the miR-29 cluster in mice, causing cardioprotection and decrease in fibrosis (123). The authors suggest that miR-29 may be more dominantly expressed in cardiomyocytes versus CFs and that whereas miR-29 may be antifibrotic in the fibroblasts, it is deleterious in cardiomyocytes, with the knockout effect dependent on the CM-specific effects.

Wingless and Int-1 (WNT) signaling is often dysregulated in fibrosis (124). miR-199a was shown to be up-regulated during fibroblast activation and to target secreted frizzled-related protein 5 in vitro. Inhibition of miR-199a reduced fibroblast migration and proliferation (125). Other studies have found that targeting fibroblast proliferation is beneficial. Cui et al. (126) showed that miR-574-5p was up-regulated in TGF-β–induced fibroblast activation. miR-574-5p targets ARID3A, which has been implicated in mediating cell cycle progression. MiR-590 targets ZEB1, an activator of transcriptional regulator CXCR4, and its overexpression inhibited migration, proliferation, and collagen secretion in CFs (127).

ROLE OF miRNAs IN ENDOTHELIAL CELL RESPONSE TO MYOCARDIAL INFARCTION

The cardiac muscle has a dense vascular network to meet the high metabolic demands of the tissue, and the continuous EC monolayer lining serves as a barrier between the blood and myocardium (128). MI severely affects these functions and induces endothelial activation, which facilitates recruitment of inflammatory cells during the inflammatory phase and mediates the repair and remodeling of the vascular network within the injured cardiac tissue via angiogenesis. Thus, miRNA therapy targeting ECs focuses mainly on regulating inflammatory recruitment and inducing angiogenesis.

Endothelial cell responses to myocardial infarction

Inflammatory recruitment

Immediately after MI, ischemia activates cardiac ECs to a proinflammatory and prothrombotic phenotype (129). The ischemic environment causes an increase in the generation of ROS and the presence of proinflammatory cytokines, such as TNFα and IL-6 (130). These conditions induce the expression of cell surface adhesion molecules, providing sites of adhesion to facilitate the recruitment and attachment of circulating leukocytes. The proinflammatory phenotype results in reduced endothelial nitric oxide (NO) production and bioavailability, impairing the ability of the endothelium to regulate vascular permeability and tone in response to external stimuli. Although endothelial activation mediates the inflammatory response within the cardiac tissue, prolonged activation and continued imbalance of ROS and NO generation can lead to permanent adverse effects, such as endothelial dysfunction and cellular apoptosis (130).

Angiogenesis

Angiogenesis, the process by which new blood vessels form from existing vasculature, is essential for cardiac repair after MI, as it is needed to restore sufficient blood flow to the injured tissue (131). Revascularization occurs at the site of the infarct as well as throughout the surrounding injured tissue via proliferation and migration of ECs. Ischemia induces expression of growth factors, which promote EC survival, migration, and proliferation, such as the vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) families (132). The hypoxic environment created during ischemia up-regulates the expression of hypoxia-inducible transcription factors (HIFs), such as HIF-1α, which activate the angiogenic process.

Recent advances in miRNA therapies targeting endothelial cells

Of the studies reviewed, relatively few focused on the EC population (24 of 213). Although some studies focus on preventing EC apoptosis or address other effects (5 of 24), most studies target the process of angiogenesis (19 of 24). Here, we summarize key miRNAs and their mechanisms of action in ECs (Fig. 4B).

Apoptosis

Apoptosis of ECs disrupts the barrier function of the endothelium and impairs its ability to effectively regulate the inflammatory response after injury. miRNA involvement in EC apoptosis has been primarily studied within the context of atherosclerotic models, which is a major cause of complications such as MI. Here, beneficial miRNAs prevent, whereas harmful ones promote EC apoptosis. The atherogenic factor oxidative low-density lipoprotein (ox-LDL) induces EC dysfunction and apoptosis and has been used both in vitro and in vivo to identify and investigate associated miRNAs. miR-26a was identified as dysregulated during atherosclerosis: Its expression was suppressed in a dose-dependent manner in human aortic ECs (HAECs) treated with ox-LDL (133). Overexpression of miR-26a targeted and repressed activity of transient receptor potential cation channel subfamily C member 6 (TRPC6), a calcium-permeable channel subunit, inhibiting apoptosis by inhibiting cytosolic calcium overload, which triggers the calcium activated apoptotic pathway.

In a separate study, ox-LDL–induced apoptosis was investigated in HAECs and human umbilical vein endothelial cells (HUVECS), and miR-150 was up-regulated during the process (134). Overexpression of miR-150 enhanced cellular apoptosis, whereas inhibition of miR-150 alleviated apoptosis. The proapoptotic miR-150 negatively regulates expression of Cu/Zn superoxide dismutase (SOD1) through the direct targeting of transcription factor ELK1. Last, miR-324 was shown to protect against hypoxia-induced apoptosis in endothelial progenitor cells (EPCs), which are active participants in the recovery process after MI, supporting both vascular endothelium repair and angiogenesis (135). Increasing expression of miR-324 down-regulates the MTFR1 gene, which, in turn, reduces mitochondrial fragmentation, stabilizes mitochondrial membrane potential, and results in decreased cellular apoptosis in hypoxic conditions.

Angiogenesis

Angiogenesis is a tightly regulated process involving multiple molecular pathways. miRNAs modulate many of the pathways involved, with beneficial miRNAs promoting and harmful miRNAs impeding the angiogenic process.

Multiple miRNAs that regulate angiogenesis modulate pathways activated by binding of VEGF (specifically VEGFA) to cell surface receptors, including HIF-1α, MAPK, and endothelial nitric oxide synthase (eNOS), which all contribute to promoting angiogenesis (132). Modulating HIF-1a is the proangiogenic miR-424, which is up-regulated in ECs under hypoxic conditions and targets Cullin 2 (136) to stabilize and up-regulate HIF-1α. In contrast, the harmful miR-223, although also up-regulated in ischemic conditions, negatively regulates HIF-1α signaling by targeting ribosomal protein S6 kinase B1 and inhibits EC migration and proliferation (137). In the MAPK pathway, the proangiogenic miR-126 and miR-132 both suppress negative regulators of MAPK activity, Spred-1 and RASA1, respectively, whereas anti-angiogenic miR-24 targets p21-activated kinase (PAK4), inhibiting MAPK activity (138–141). miR-24 also targets the eNOS signaling pathway, along with miR-199a and miR-155, all of which suppress migration, proliferation, and tube formation of ECs (142, 143). Inhibition of these miRNAs increases NO bioavailability and promotes angiogenesis in these studies. Although these miRNAs target molecules of pathways activated by the binding of VEGF, certain miRNAs target VEGF directly, such as the antiangiogenic miR-590 (144).

Several other miRNAs affect angiogenesis by modulating signaling pathways that are not VEGF associated. In two separate studies, miR-210 was shown to promote angiogenesis, with overexpression increasing EC proliferation and migration by up-regulating hepatocyte growth factor expression and targeting focal adhesion kinase (145, 146). miR-92 also showed, in two separate studies, that it targets the proangiogenic integrin α5 (ITGF5) and that its inhibition substantially increased angiogenesis and granulation tissue formation (147, 148). Because miR-26a targets SMAD1, its inhibition led to increased angiogenic activity and reduced infarct size in a murine model of MI. In contrast, miR-27b promotes vascularization by targeting notch ligand DLL4 (149). Last, miR-185 inhibits proliferation, migration, and tube formation by targeting the CatK gene (150). Inhibition of miR-185 significantly promoted these functions of angiogenesis under hypoxic conditions.

Other effects

Bayoumi et al. (151) determined that miR-532 overexpression decreased endothelial-to-mesenchymal transition (EndMT) in cardiac ECs. miR-532, which is up-regulated by β2 adrenergic receptor/β-arrestin activity, exerts this cardioprotective function by targeting protease serine 23 (PRSS23), a positive regulator of maladaptive EndMT. The metabolic state of ECs can also alter MI outcomes, as Bartman et al. (152) identified that miR-21 increased glycolytic activity in ECs in hypoxic conditions. miR-21 was up-regulated under hypoxic conditions only in ECs, not in cardiomyocytes or fibroblasts, despite being predominantly expressed in fibroblasts under normoxic conditions. Inhibition of miR-21 resulted in larger infarct sizes in a myocardial IRI murine model.

ROLE OF miRNAs IN IMMUNE CELL RESPONSE TO MYOCARDIAL INFARCTION

MI triggers an inflammatory cascade in the heart starting with an acute proinflammatory response, which is gradually replaced by an anti-inflammatory reparative phase. Therapeutic advancements generally aim to dampen the initial inflammatory response and promote the proreparative phase.

Inflammatory responses to myocardial infarction

Activation of inflammation

Upon injury, resident immune cells located in perivascular areas are activated, resulting in an acute inflammatory response. In the early inflammatory phase, the infiltration of immune cells in the infarct area and accumulation of proinflammatory chemokines/cytokines result in not only the clearance of cellular debris but also in further damage to cardiac tissue. Macrophages, among the largest resident cell populations in cardiac tissue, are predominantly responsible for the production of inflammatory cytokines post MI. They release IL-1, IL-6, and TNFα, triggering further responses from multiple cell types, both locally and remotely. The activation of this inflammatory cascade mediates remodeling throughout the myocardium.

Resolution of inflammation

Increasing evidence indicates that the initial inflammatory response elicited by MI induces resident macrophages to switch from an inflammatory M1 phenotype to a resolving M2 phenotype. M2 macrophages secrete anti-inflammatory signals critical for the repair response, allowing wound healing and scar formation, thereby limiting infarct size. Failure to activate reparative responses leads to adverse cardiac remodeling with further damage and, ultimately, heart failure.

Recent advances in miRNA therapies targeting immune cells

Of the studies surveyed, 17 of 213 focus on immune cells. Although some studies have applied miRNA therapies to either reduce the damage of the inflammatory response (10 of 17 studies) or focus on promoting the reparatory pathways (3 of 17 studies), several other effective therapies have targeted both responses, exerting the overall functional benefits as an ensemble (4 of 17 studies). In targeting immune cells, beneficial miRNAs aim to prevent infiltration of immune cells and inflammatory signal production in the infarct area and promote M2 macrophage polarization, efferocytosis, phagocytosis, and angiogenesis signal release. Here, we summarize key miRNAs and their mechanisms of action in immune cells (Fig. 4C).

Infiltration

Infiltration of immune cells such as macrophages and neutrophils into the infarct area is one of the initial inflammatory responses after MI. Many miRNAs have been shown to regulate this process, providing interesting therapeutic targets. Recently, miR-141 has been shown to suppress neutrophil/leukocyte adhesion to ECs by modulating expression of the cell adhesion associated protein intercellular adhesion molecule–1, which plays a crucial role in regulating the migration of leukocytes across the endothelium into the myocardium. In vivo, pretreating mice with intravenous miR-141 mimics before IRI successfully reduced CD11b⁺ myeloid cells and F4/80⁺ macrophages accumulation in the ischemic myocardium (153).

miR-21 knockout mice showed enhanced infiltration of CD11b⁺ monocytes/macrophages in myocardium, whereas miR-21 overexpression markedly inhibited it (154). However, contrasting findings in another study showed that miR-21 inhibition in cardiac allografts led to reduced infiltration of CD45⁺ leukocytes, but there was no difference observed for macrophage infiltration (155). Intravenous injection of miR-144 mimics resulted in reduced infiltration of macrophages in the border zone, decreased MMP-mediated inflammation, and reduced fibrosis and apoptosis. The beneficial effects of miR-144 therapy were associated with mTOR- and P62-mediated autophagy signaling with improved cardiomyocyte survival. Alterations in TGF-β signaling also demonstrated a role for miR-144 in modulating the local inflammatory response (156).

Intramyocardial injection of anti–miR-224–expressing lentivirus elevated vascular cell adhesion molecule–1 expression, regulating inflammation-associated vascular adhesion and transendothelial migration of macrophages/leukocytes, and MMP-2, an inflammatory mediator participating tissue remodeling in pathological processes (157). Other studies have shown that the overexpression of miR-106b, miR-148b, and miR-204 injection of miR-181b–enriched cardiosphere-derived exosomes (CDC-exo) or miR-24–enriched umbilical mesenchymal stem cells (MSCs) secreted exosomes (MSC-exo), and knockdown of miR-375 all suppressed CD68⁺ cell accumulation in the infarct border zone, leading to improved cardiac function with decreased infarct size (158–161).

Inflammatory cytokine production

The accumulation of proinflammatory cytokines plays a critical role in governing the progression and extent of tissue remodeling. The inhibition of this process by miRNA therapies successfully decreases myocardial infarct size and improves cardiac function. miR-21 mimic delivered to monocytes/macrophages in mice reduced inflammatory cytokine expression by directly inhibiting KBTBD7-mediated DAMP (damage-associated molecular patterns)–triggered inflammatory responses in macrophages (154). The delivery of miR-106b, miR-148b, and miR-204 individually encapsulated into polyketal nanoparticles in macrophages decreased Nox2 expression, inhibiting superoxide production and expression of proinflammatory genes IL-1α, IL-6, and TNF-α in vitro (158). Anti–miR-375 injection after MI in mice decreased inflammatory response by reducing the production of several proinflammatory cytokines including IL-6, interferon gamma-induced protein 10 (IP-10), monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein-1 alpha (MIP-1α), regulated on activation, normal T cell expressed and secreted (RANTES), and MIP1-β (160). After reperfusion in rats and swine, intramyocardial injection of miR-181b–enriched CDC-exo also attenuated the expression of proinflammatory genes by targeting PKCδ (159).

M2 macrophage polarization

Polarization of macrophages toward the anti-inflammatory M2 phenotype has been found to confer beneficial effects on cardiac repair. Intramyocardial injection of MSC-exo containing miR-182 after myocardial IRI reduced infarct size and alleviated damaged myocardial inflammation in mice. Mechanistically, miR-182 polarized macrophages toward the M2 phenotype and regulated downstream Toll-like receptor 4 (TLR4) expression, mitigating proinflammatory cascades and enhancing subsequent repair (162).

Tail vein injection of miR-21 mimic delivered to cardiac macrophages by encapsulated nanoparticles promoted macrophage switching from a proinflammatory to reparative phenotype, thus inducing the resolution of inflammation (163). The polarization state was also shifted in cardiac macrophages isolated from CDC-exo–treated rats and swine. miR-181b inhibition of PKCδ expression further demonstrates that PKCδ acts as a downstream effector of CDC-exo–mediated cardioprotection (159). Moreover, increased CD206⁺ cells (M2 macrophage marker) in the left ventricular infarct border zone of anti–miR-375–treated mouse hearts after MI indicate that miR-375 knockdown can trigger a macrophage switch toward the anti-inflammatory M2 phenotype (160). In vitro studies in macrophages further demonstrated that miR-375 modulated AKT signaling through phosphoinositide-dependent kinase-1 (PDK-1). PDK-1 knockdown abolished the macrophage polarization effect of miR-375, further confirming that PDK-1 is a critical mediator in miR-375–regulated macrophage phenotype shift.

Targeting efferocytosis and phagocytosis

The clearance of apoptotic cells by macrophages is thought to play a critical role in the reparative phase allowing for further recovery after MI. MI-associated transcript (MIAT), a long noncoding RNA, has been shown to regulate phagocytosis by acting as a sponge of miR-149, which directly targets antiphagocytic molecule CD47 (164). Knockdown of MIAT or miR-149 overexpression in murine macrophages reduced CD47 expression and promoted efferocytosis and phagocytosis, which consequently increased resolution of inflammation and clearance of apoptotic cells by macrophages. Similarly, transfecting EPCs derived from the peripheral blood of patients with ST-segment elevation MI with miR-324 mimic resulted in reduced apoptosis, increased proliferation, and elevated phagocytosis by targeting and inhibiting MtFr1 after peroxide-induced oxidative stress (135). The protective role of miR-324 overexpression against oxidative stress-induced EPCs injury suggests that therapies leading to correction of miRNA expression may contribute to tolerance against injury by regulating phagocytosis pathways.

Targeting angiogenesis signal release

Macrophages engage in cross-talk with other cell types and release a variety of angiogenic signals to initiate and regulate angiogenesis in ECs in response to MI. After MI, M1 macrophages secrete proinflammatory exosomes (M1-exo), which contain abundant miR-155 (142). M1-exo containing miR-155 suppressed expression of several genes, including RAC1, PAK2, Sirt1, and AMPKα2, and modulated Sirt1/AMPKα2-eNOS and Rac family small GTPase 1 (RAC1) - p21 (RAC1)–activated kinase 2 (PAK2) signaling pathways, leading to the inhibition of angiogenesis and cardiac dysfunction. The antiangiogenic function of miR-155 makes its inhibition an attractive therapeutic strategy for MI (142).

miRNAs WITH PLEIOTROPIC EFFECTS

Most miRNA studies focus on one miRNA in a singular cell type. However, the literature demonstrates that many miRNAs have pleiotropic effects on myocardial recovery after MI, spanning different pathways, processes, and cell types. Some pleiotropic miRNAs regulate many points along the same pathway, whereas others regulate different pathways. Interestingly, multiple of these miRNAs demonstrate divergent effects among cell types—promoting recovery processes in some cell types while impeding recovery processes in other cell types. Developing a clear mechanistic understanding of the pleiotropic effects of miRNAs will greatly advance their clinical translation.

miRNAs with convergent effects

Some miRNAs have convergent effects: promoting or impeding recovery processes in all cell types. For example, miR-210, miR-324, miR-494, and miR-532 all have beneficial effects on both cardiomyocytes and ECs (42, 135, 145, 146, 151, 165–169). Although miR-324 had the same target (MTFR1) and effect (prevention of apoptosis) in both cell types (135), the other three miRNAs had different targets and affected different pathways in each cell type. In separate studies, miR-24 was shown to have beneficial effects on cardiomyocytes, fibroblasts, and ECs through the prevention of apoptosis, prevention of fibrosis, and activation of angiogenesis, respectively (24, 99, 141). On the other hand, miR-92a has been shown to have harmful effects on fibroblasts and ECs through the activation of fibrosis and the prevention of angiogenesis (105, 147, 148). Because of their convergent effects on multiple cell types studied, these miRNAs may be particularly promising therapeutic targets.

miRNAs with divergent effects

Interestingly, most of the pleiotropic miRNAs identified demonstrate divergent effects: promoting recovery processes in some cell types while impeding recovery processes in other cell types. We already described miR-29, which demonstrated contradictory effects on CFs versus cardiomyocytes. Another extensively studied miRNA that falls into this category is miR-199a. miR-199a was shown to induce adult cardiomyocyte proliferation after MI (70), and subsequent studies identified multiple proproliferative pathways regulated by miR-199a in cardiomyocytes (71). Additional studies have demonstrated that miR-199a can regulate other crucial pathways in cardiomyocytes including glucose metabolism, hypertrophy, apoptosis, and autophagy with divergent effects (66, 170–173) and negatively affect angiogenesis in ECs as well as induce pathological fibrosis in fibroblasts (125, 143).

Another noteworthy miRNA is miR-21. Multiple studies have demonstrated that miR-21 prevents cardiomyocyte apoptosis after ischemic injury through direct inhibition of the proapoptotic molecule PDCD4 (38–40). In addition, treatment with miR-21 mimics has been beneficial to ECs by preventing apoptosis (152) and to immune cells by preventing inflammatory activation (152, 154, 163). However, despite consensus on the beneficial effects of miR-21 in cardiomyocytes and ECs, multiple studies have demonstrated harmful effects of miR-21 in fibroblasts, and other studies have shown contradictory effects in immune cells. The knockdown of miR-21 prevents pathological fibrosis by targeting the ERK/MAPK and TGF-β pathways (101, 111–113). In immune cells, miR-21 has been shown to both prevent macrophage infiltration and induce leukocyte infiltration (154, 155). Ultimately, despite contradictory effects, the strongest evidence for the therapeutic capabilities of miR-21 comes from a recent study by Hinkel et al. (174), which demonstrated that inhibition of miR-21 prevented myocardial dysfunction in a pig model of MI.

Multiple miRNAs have been shown to have divergent effects on fibroblasts compared to cardiomyocytes. For example, miR-22, miR-29, miR-101, and miR-155 are all beneficial to fibroblasts by preventing activation and proliferation while being harmful to cardiomyocytes by promoting apoptosis or autophagy (59, 96, 97, 100, 120, 142, 175–179). In contrast, miR-144 prevents cardiomyocyte apoptosis (beneficial) while also promoting fibroblast proliferation (harmful), by targeting the same PI3k/AKT pathway in both cell types (156). Collectively, these miRNAs demonstrate the complexity of miRNA effects across multiple cell types in the heart and suggest the difficulties with using nontargeted miRNA therapies. Thus, a cardiomyocyte-centric approach to miRNA therapy limits our understanding of pleiotropic effects.

PROSPECTS AND CHALLENGES FOR CLINICAL TRANSLATION

Promising preclinical studies have spurred interest in clinical translation of miRNA therapies. Currently, there are no miRNA therapies with U.S. Food and Drug Administration (FDA) approval; however, several early-stage biotechnology companies are focused solely on miRNAs, such as Miragen Therapeutics, Cardior Therapeutics, and Regulus Therapeutics. One phase 1 trial is currently ongoing to assess the safety of a miR-132 inhibitor (NCT04045405) in ischemic heart disease. There are several clinical trials for treatment of related pathologies, including the use of a miR-21 inhibitor for prevention of renal fibrosis (NCT03373786) and a miR-29 mimic for pulmonary fibrosis (NCT03601052).

As discussed here, mimicry of “beneficial” miRNAs or inhibition of “harmful” miRNAs both have therapeutic potential. However, the latter has benefited from past innovation in the RNA interference field, which has recently begun, achieving success in the clinic, with the first drug of its class, Patisiran, approved by the FDA in 2018. It is thus not surprising that of the current miRNA therapeutics in either phase 1 or 2 clinical trials, the majority are ASOs targeting a specific miRNA (180), although mimics are also being tested. Similar to the siRNA field, these burgeoning new therapies face challenges including considerations for therapeutic timing, optimal oligonucleotide optimization, in vivo delivery, and side effects.

Oligonucleotide modifications

Naked oligonucleotides rapidly accumulate in the kidney and liver and are quickly cleared from the circulation (181). Moreover, cleavage by serum exonucleases and degradation in the intracellular endosomal compartment reduces drug potency. To improve stability, various chemical modifications to the nucleic acid backbone have been used in the design of miRNA mimics and inhibitors.

Phosphothiorates substitute a sulfur for oxygen in the phosphate group of the nucleotide. Compared to unmodified oligonucleotides, phosphothiorates are more resistant to nucleases and thus drastically increase circulation time (182). However, compared to an unmodified phosphodiester bond, they have decreased binding affinity to their target as measured by a lower melting temperature (T_m) (183). To address these issues, second-generation designs centered around the use of 2′-O modifications of the ribose sugar and locked nucleic acid (LNA) modifications while reducing and interspersing the number of phosphothiorate modifications. 2′-O-methylation (2′OMe), first tested in the early 2000s, enhances target binding (184). Krützfeldt et al. (185) used 2′OMe oligonucleotides with spaced phosphothiorate bonds and a cholesterol group at the 3′ end to increase resistance to exonucleases while preserving nuclease activity, so termed “antagomirs” (186).

LNA modifications contain methylene linkages of the 2′O to the 4′C in the sugar backbone, locking the structure into a C3′-endo sugar conformation (186). These afford an increase in T_m while simultaneously increasing nuclease resistance (187). Given their high binding affinity, much shorter LNA oligonucleotides can be designed. This has both implications for target specificity, as shorter sequences more likely bind multiple families of miRNAs, and the related cost of synthesizing such an oligonucleotide. Recently, Obad et al. (188) reported the use of an 8-mer LNA oligo capable of powerful anti-miRNA effects.

Last, miRNA sponges have been explored as alternatives to ASOs (189). These dominant-negative inhibitors contain multiple target sites complementary to a specific miRNA seed sequence. They have several advantages over traditional oligonucleotides, including the ability to silence families of RNA sharing a common seed. However, further progress on their design and delivery is required for clinical translation, and this area of research still trails ASOs in terms of optimization (190).

Delivery vehicles

The route of administration of miRNA therapeutics greatly alters their potency. Local administration methods such as intracoronary injection (IC), hydrogel-based patches, and intramyocardial injection have been studied (table S1). The availability of percutaneous coronary intervention as the preferred method for revascularization after MI (191) makes IC injection perhaps the most effective local delivery method. Despite this, the multiple dosing of locally administered drugs to the heart is invasive and difficult in humans. In a preclinical study of MI in 135 swine, Foinquinos et al. (192) demonstrated the efficacy of IC injections of miR-132 immediately after MI. Notably, they also show that intravenous infusions were as effective as IC injections, suggesting the possibility for minimally invasive delivery of miRNA therapies.

Intravenously injected ASOs delivered without a vehicle still require large doses to be effective in vivo (185). In addition, efficient delivery to the target tissue—in this case, the heart—is essential for optimizing in vivo delivery of miRNA therapies. For this purpose, carriers such as lipid nanoparticles (LNP) and nanoparticles (193) have been explored. Traditional LNPs have been clinically validated for siRNA delivery, including Patisiran. However, their use comes at the cost of additional components and toxicities. Because these systems have been reviewed extensively (181, 194), they will not be discussed in detail here.

Extracellular vesicles, including exosomes, are plasma membrane–bound cell-secreted vesicles that transfer biologically active cargo between cells and act as the endogenous analog of LNPs. In the field of heart repair, exosomes emerged first as paracrine mediators of cell therapy. They are secreted by all cells and contain a variety of contents, including miRNAs (195). Exosomes from several cell types, including MSCs, cardiac progenitor cells, and induced pluripotent stem cell–derived cardiomyocytes (iPS-CMs), are effective in promoting cardiac repair (196–198). Several advances have been made toward clinical translation, including the development of Good Manufacturing Practice protocols for the collection of therapeutic exosomes (199) as well as toward altering their cargo. Unlike transplanted cells and synthetic LNPs, autologous exosomes do not trigger the immune system. However, barriers to their use remain, like addressing their heterogeneity and production. Eventually, highly defined engineered exosomes may represent a way to fine-tune miRNA therapy for the heart. For example, MSCs overexpressing miR-133 were more effective in cardioprotection than control vector MSCs in a rat model of MI, an effect attributed to an increase in therapeutic exosome content (200).

Tissue-specific delivery

At this time, targeted delivery of miRNA therapeutics to the heart remains an unmet need. Biodistribution studies show that LNP and exosomes are entrapped by the liver and spleen and filtered from the blood by the kidneys’ glomerular barrier due to their size (201). With only an estimated 2000 to 5000 copies of oligonucleotide required within the cell for gene knockdown, drug loading doses can be reduced if delivery can be constrained to the heart (202). The conjugation of ASOs to N-acetylgalactosamine (GalNAc) to specifically target asialoglycoprotein receptors in the liver has contributed greatly to the commercial success and translation of Patisiran (203). Identifying similar extrahepatic targets, including in the heart, has not yet been achieved and is a key focus.

To this end, homing cardiac targeting peptides and engineered exosomes have been shown to increase exosome and miRNA delivery in cardiomyocytes after IC injection (204). Chemical aptamers can alter biodistribution. Xue et al. (205) used a nanoparticle dendrimer approach to target miR-1 inhibitors to the AngII receptor 1 in post-MI hearts. Antibody-based targeting can also alter cardiac homing. Immunoglobulins are immunologically privileged proteins and are continually recycled in the circulation. Conjugation of siRNAs to an anti-CD71 Fab′ fragment allowed for durable gene silencing in cardiac tissues over a month’s time (206). Liu et al. (207) used anticardiac troponin antibodies to target liposomes containing anti–miR-1 to ischemic myocardium. However, the addition of a large macromolecule may also lead to several disadvantages. Genentech’s THIOMAB antibody-siRNA conjugates demonstrated the ability to specifically target prostate cancer in vivo but exhibited suboptimal gene silencing due to sequestering in the endocytic department (208). In addition, targeting specific receptors may have other effects besides delivering the therapeutic molecule, such as inhibition or activation of its own downstream signaling pathway, which adds complexity to this delivery modality.

Endosomal escape

After organ-specific targeting, there remains a barrier to entry at the cellular level. miR-mimics and anti-miRs are relatively large (~10,000 kDa) and highly negatively charged, preventing membrane transport. Naked nucleic acid or those bound in carriers are ubiquitously taken up by endocytosis (209). Endocytosed molecules then enter the endosomal system and eventually are degraded in the lysosome (210). To have a clinical effect, miRNA therapeutics must escape the endosomal pathway into the cytosol. Using traceable siRNAs, Gilleron et al. (211) estimated that just 1% of delivered oligonucleotides actually escape the endosomal pathway. The exact biological mechanisms of how escape occurs remain to be elucidated. Approaches to designing delivery carriers, which bypass the endosomal system via diffusion through the endosomal membrane or disrupting the compartment pH, have been explored (212). Solving the endosomal escape barrier will be critical for miRNA therapeutic development moving forward.

Side effects of miRNA therapies

Like any drug, miRNA therapeutics are not without safety concerns. For example, phosphothiorates have also been reported to react with proteins, including FGF2 (213). Although careful sequence selection can eliminate off-target RNA hybridization, many oligonucleotide modifications produce nonspecific effects, including differences in protein expression when compared to an unmodified siRNA (214).

Nucleic acids also broadly trigger the immune system, activating the TLR family (215). The use of delivery vehicles such as LNPs, which are designed to shield ASOs from the immune system, adds further reagents that may trigger an inflammatory response. A miRNA-34a mimic was halted in phase 1 (NCT02862145) studies after several patients developed severe adverse immune reactions. Development of these therapeutics can be thus unpredictable.

Abrogating the acute response to MI may also be dangerous. Therapies may be able to salvage damaged but viable cells adjacent to the infarct; however, an injured cardiomyocyte that is rescued may have impaired function, leading to arrhythmia. Similarly, interference of fibroblast function after MI has long been controversial, due to fear of cardiac rupture after failure to generate a scar. Although no studies reviewed here reported such outcomes, developing clear understandings of the mechanisms underlying miRNA therapies will go a long way toward preventing these dangerous side effects as we continue to search for more potent miRNA therapies.

The ability of miRNAs to regulate a variety of processes in different cell types leads to the additional concern of perturbations in signaling outside of the target organ. Although promoting cardiomyocyte proliferation may be desirable for cardiac recovery, altering those same pathways in other organs may result in tumorigenesis. The aforementioned miR-199a, which caused uncontrolled cardiomyocyte proliferation and subsequent arrhythmia, has also been linked to carcinogenesis and metastasis in certain cancers, such as melanoma and gastric cancer (216, 217). Similarly, inhibition of miR-34a has been shown to improve cardiac remodeling after injury (74, 218), although, at the same time, miR-34a has been studied as a tumor suppressor, and its mimicry has demonstrated anticancer effects (219, 220). Collectively, these examples further highlight a need to understand cell type–specific effects of miRNAs as well as the need for tissue-specific delivery.

Alternative preclinical models

Another challenge in miRNA therapeutic development, similar to genome-based therapies, is the lack of adequate drug-testing platforms. Genomic differences in nonprimate animal models limit their use as a predictive model of the effects of a specific miR-mimic or anti-miR. Moreover, interspecies differences in ion channels and biological pathways fail to recapitulate human biology; for example, the mouse heart beats nine times faster than the human heart (221). Current cardiovascular drug testing relies heavily on simplified models such as human embryonic kidney 293 cells and Chinese hamster ovary cells overexpressing ion channels (222). Of the studies surveyed here, most used primary rat cardiomyocytes or the rat H9C2 cell line, highlighting a need for in vitro human disease models. Although primary human cardiomyocytes would be an ideal candidate for such a model, they are in prohibitively short supply and difficult to isolate. Recently, iPS-CMs have emerged as a promising source to fill this void (223). iPS-CMs can be produced in near limitless quantity and matured to adult-like phenotypes (224), allowing for high-throughput screening of thousands of miRNA candidates. The generation of more complex, three-dimensional models of human cardiac tissues using iPS-CMs can further be used to model MI (225). Engineered cardiac tissue can also help shed light on miRNA transfer between cardiac cell types and myocyte-nonmyocyte communication, through extracellular vesicles or otherwise (226). Elucidating this biology may help further the design and development of miRNA therapeutics. How closely these models recapitulate human biology remains to be seen (227); however, they represent exciting technologies that can potentially decrease drug development cost.

Key remaining questions

There are several questions, which, when elucidated, would greatly advance the field of miRNA therapy for MI. A key question that remains is the timing of therapeutic delivery. The processes of cardiomyocyte death, fibroblast activation, and immune cell response occur acutely after ischemia onset. Because of the acute nature of cardiac injury, is it then necessary for intervention to take place rapidly after symptom onset? Or can a treatment administered in the later stages of wound healing (3 to 4 weeks) still show clinical improvement? Along a similar vein, could miRNA therapies potentially be preventative, suppressing these maladaptive processes before they begin? Very few preclinical studies have examined the timing of these deliveries and instead apply both the insult and the treatment simultaneously. Furthermore, the potential duration of clinical benefit from miRNA therapy is not well defined. Given the short-acting nature of miRNAs, it seems that repeated dosing would be required for long-lasting clinical benefit if early intervention is not possible. Should we target one cell type’s pathobiology or multiple? Is there a dominant effect of manipulating one cell type over another? The pleiotropic nature of miRNAs with convergent and divergent effects is one of its key advantages, yet it adds a degree of complexity to miRNA drug discovery unparalleled in other treatment modalities. The recent advancements in human iPS-derived in vitro models, including the ability to generate complex, three-dimensional models of human cardiac tissues, may represent a powerful and efficient strategy in answering some of these challenging questions.

CONCLUSION

Current interventions fail to recover the injured myocardium after MI, requiring the development of previously unexplored cardioprotective strategies. Here, we reviewed miRNAs as powerful regulators involved in post-MI remodeling and as the potential targets for MI therapy. Focusing on literature within the past 3 years and the foundational studies within the past 10 years, we identified a total of 213 publications describing the modulation of miRNAs in the treatment of MI, involving 116 unique miRNAs. Most of these miRNAs have been shown to play both beneficial and harmful roles in the progression of MI with divergent effects on different target cell types. To successfully translate the miRNA-based therapies of MI to the clinic, a more detailed understanding of the mechanisms and dynamics of miRNA effects on each cell type needs to be clearly defined. Clinical translation of these therapies will also require improved administration protocols, organ-specific delivery, and more predictive preclinical models.

Supplementary Material

Supplemental table 2

Table S2. Summary of miRNAs reviewed.

Supplemental table 1

Table S1. Complete list of miRNAs reviewed.