MiRiad Roles for MicroRNAs in Cardiac Development and Regeneration

Cardiac development is an exquisitely regulated process that is sensitive to perturbations in transcriptional activity and gene dosage. Accordingly, congenital heart abnormalities are prevalent worldwide, and are estimated to occur in approximately 1% of live births. Recently, small non-coding RNAs, known as microRNAs, have emerged as critical components of the cardiogenic regulatory network, and have been shown to play numerous roles in the growth, differentiation, and morphogenesis of the developing heart. Moreover, the importance of miRNA function in cardiac development has facilitated the identification of prospective therapeutic targets for patients with congenital and acquired cardiac diseases. Here, we discuss findings attesting to the critical role of miRNAs in cardiogenesis and cardiac regeneration, and present evidence regarding the therapeutic potential of miRNAs for cardiovascular diseases.


Introduction
The survival of any metazoan during embryonic development and adult life depends on the continuous function of the heart. Accordingly, congenital heart abnormalities, which are estimated to affect approximately 1% of newborns [1][2][3], account for the greatest proportion of birth defect-related MiRNAs are typically transcribed in the nucleus by RNA polymerase II (RNA Pol II), and can be encoded within protein-coding exons, the introns of protein-coding genes, or intergenic regions. Following transcription, the capped, poly-adenylated primary miRNA (pri-miRNA) is cleaved by the RNAse III-like enzyme, Drosha (Pasha in flies), and its partner, DGCR8, to generate a ~70 nucleotide-long pre-miRNA molecule that contains local stem-loop secondary structure. The pre-miRNA is then exported to the cytoplasm by the Ran-GTP-dependent nuclear transport receptor, exportin 5, where the terminal loop is cleaved by the RNAse III-like enzyme, Dicer. The resulting double-stranded ~22 nucleotide-long miRNA duplex consists of a "guide" and "passenger" strand; the "guide" strand is preferentially loaded into the RISC (RNA-induced silencing complex) as the mature miRNA, and ultimately guided to its target mRNA transcript.

Components the miRNA Biogenesis Pathway Are Required for Cardiac Development
One approach to evaluating the requirement for miRNAs during cardiac development has been to genetically manipulate Dicer, the enzyme responsible for processing pre-miRNAs into their mature forms. In vertebrates, this protein is encoded by a single gene, Dicer1. In mice, null mutations in Dicer1 are lethal by approximately E7.5 (embryonic day 7.5), as a failure to process pre-miRNAs to maturity may result in the depletion of pluripotent stem cells (Table 1) [42]. Similarly, Dicer1-null zebrafish embryos exhibit generalized growth arrest, and typically do not survive beyond 13-14 days post-fertilization (d.p.f.) [43].  [53,54] Mouse Null: neomycinresistance cassettes retained Lethal (P10): VSD, aortal misalignment; cardiac dilation; sarcomere disruption and retention of fetal sarcomere gene expression program [52] Null: neomycinresistance cassettes excised Lethal (P17): Dilative cardiomyopathy, CM hyperplasia; retention of fetal sarcomere gene expression program [55]  Due to the severity of global Dicer1 loss-of-function mutations, tissue-specific deletion models have become essential to more completely understand the role of this enzyme in cardiac development. A conditional knockout of Dicer1 using Cre recombinase under the control of endogenous Nkx2.5 (Nk2 homeobox 5) regulatory regions (Nkx2.5-Cre) ablated Dicer activity in cardiac progenitor cells beginning at E8.5 [44]. Embryos lacking Dicer1 in the heart died by E12.5, and exhibited a poorly developed myocardium and pericardial edema. Interestingly, however, many key regulators of cardiac differentiation and patterning, such as Hand2 (heart and neural crest derivatives expressed), Tbx5 (T-box 5), and Mlc2v (myosin light chain 2v), were all expressed normally [44]. A second gene-targeted Nkx2.5-Cre line that made use of a different Cre allele (3'UTR-IRES-Cre) than the one previously employed by Zhao et al. [44] was subsequently used to reveal the requirement of Dicer for septation of the cardiac chambers and alignment of the outflow tract [45]. These embryos survived until approximately E13.75, slightly longer than the mutants generated in the preceding study. The extended survival of these mice was important because it permitted Tabin's group to study the effects of Dicer1 loss of function on morphological events that occurred too late during development to evaluate with the alternate Nkx2.5-Cre [45]. Taken together, although these two Nkx2.5-Cre lines exhibited slightly different spatiotemporal expression kinetics and manifested considerable phenotypic variation, the results of these studies indicate a requirement for cardiac-specific Dicer1 expression in the development and morphogenesis of the heart.
Neural crest cells (NCCs) arise from the dorsal neural tube during early development, and differentiate to give rise to the peripheral nervous system, melanocytes, and craniofacial tissues. A subset of these cells, known as cardiac NCCs, are essential for various aspects of cardiac morphogenesis, including outflow tract septation, as well as development of the valves, the cardiac conduction system, and the pharyngeal arch arteries [46]. To investigate the effects of loss of Dicer activity in NCCs, Zehir et al. used a Wnt1-Cre line to conditionally delete Dicer1 in this cell population. Although this mutation had no effect on NCC migration or target tissue colonization, NCC-specific deletion of Dicer1 resulted in a loss of craniofacial bones, as well as enteric, sympathetic, and sensory neurons. Moreover, caspase-dependent apoptotic pathways were activated in the developing sensory ganglia and the sympathetic nervous system, indicating that Dicer activity is crucial for the survival of certain NCC-derived tissues [47]. A subsequent study with the same Wnt1-Cre line confirmed the requirement for Dicer in craniofacial development, but further demonstrated that Dicer loss of function did not affect the survival of tissues derived from cardiac NCCs. However, these mutant mice displayed many cardiovascular defects reminiscent of human congenital heart abnormalities, including type B interrupted aortic arch, double outlet right ventricle, ventricular septal defect (VSD), retroesophageal right subclavian artery, and ectopic placement of the carotid arteries [69]. Similar defects were observed in conditional, NCC-specific Dgcr8 knockouts [50]. The effects of NCC-specific Dicer ablation were shown to be regulated in part by miR-21 (miRNA-21) and miR-181a-mediated upregulation of Erk1/2 (extracellular signal-regulated kinase) signaling [69]. Importantly, however, aberrations in cardiac development have not been noted for either miR-21 [70] or miR-181a knockout mice [71]. Therefore, studies that seek to characterize miR-21/miR-181a double knockout animals are required to further evaluate the role of these two miRNAs in NCC lineage specification. Alternatively, other miRNAs might be involved in NCC development that compensate for the functions of miR-21/miR-181a in the knockout mice.
In addition to its critical requirement during gestation, Dicer activity is also necessary for maintaining post-natal cardiac function and structural integrity. In contrast to the Nkx2.5-Cre lines, cardiomyocyte-specific Dicer ablation using an αMHC (α-myosin heavy chain)-Cre line was not embryonic lethal. However, dramatic reductions in the levels of mature miRNA in neonatal hearts were associated with features of dilative cardiomyopathy and heart failure, which resulted in death by post-natal day 4 (P4) [48]. To assess the requirement of Dicer activity in the hearts of older mice, Cre recombinase was expressed under the control of a tamoxifen-inducible αMHC promoter [49]. Cre-mediated excision of Dicer1 in three week-old mouse hearts led to spontaneous cardiac remodeling, as evidenced by atrial enlargement and mild inflammation. These morphological changes resulted in a rapid loss of cardiac function, and high rates of sudden death by four weeks of age [49]. Spontaneous cardiac remodeling was also observed in adult mice harboring cardiomyocyte-specific Dicer1 deletions. Unlike the younger mice, adult animals did not display premature sudden death. However, they exhibited more severe histopathological changes, including cardiomyocyte hypertrophy and disarray, extensive inflammation, and interstitial fibrosis [49]. A role for miRNA processing in the maintenance of cardiac function is further supported by muscle-specific deletion of Dgcr8. When expressed under the control of the MCK (myosin creatine kinase) promoter, Cre recombinase is activated in cardiac and skeletal muscle at birth, and declines to 40% of peak levels by P10 [51]. As with Dicer1-mutant mice, deletion of Dgcr8 using MCK-Cre markedly reduced the thickness of the ventricular myocardium, and disrupted the cardiac conduction system. These severe perturbations in cardiac function ultimately progressed to dilated cardiomyopathy, and caused premature lethality by two months of age [51]. Together, these studies highlight the essential role of the miRNA biogenesis pathway in the maintenance of cardiac function and integrity during neonatal and adult life. The subsequent sections of this review will examine the functions of specific miRNAs and miRNA families in the regulation of cardiac development. Although recent findings have also uncovered a role for miRNAs in vasculogenesis and vascular biology, many comprehensive reviews on this topic have been published elsewhere [72][73][74][75], and will not be discussed here.

The mIR-1/133a Bicistronic Clusters Are Critical Regulators Cardiac Development
miR-1 is an evolutionarily conserved miRNA that is highly enriched in cardiac and skeletal muscle [76][77][78]. Indeed, miRNA sequencing has revealed that miR-1 is the most abundant miRNA in the mammalian heart, and accounts for approximately 40% of all miRNA expression in that tissue [51]. miR-1 is transcribed as part of a bicistronic cluster together with miR-133a, a closely related miRNA that is also enriched in striated muscle [77]. The vertebrate genome contains two copies of the miR-1/133a locus, likely due to a gene duplication event [78]. Although these loci are located on different chromosomes (Figure 2), the mature miR-1 (miR-1-1 and miR-1-2) and miR-133a (miR-133a-1 and miR-133a-2) paralogs display complete sequence identity. An additional miRNA cluster consisting of miR-206 and miR-133b is homologous to miR-1, but is expressed uniquely in skeletal muscle ( Figure 2) [79].
Muscle-specific expression of the miR-1 and miR-133a family members is tightly controlled by a number of critical myogenic transcription factors, including SRF (serum response factor), Mef2 (myocyte enhancer factor-2), MyoD, myogenin, and Nkx2.5 [76,[78][79][80]. In mice, mutation of predicted SRF binding sites within the mIR-1-1 and miR-1-2 enhancer regions disrupted miR-1 expression in the heart, while cardiac-specific deletion of Srf ablated expression of the miR-1 transcript [76]. In addition, Liu et al. [78] observed that expression of both miR-1/133 bicistronic clusters was reduced in the hearts of Mef2C; Mef2D double knockout mice. Accordingly, they identified a muscle-specific, MEF2C-dependent enhancer located between the miR-1-2 and miR-133a-1 coding regions that directed miR-1-2/133a-1 expression in the embryonic ventricular myocardium [78]. A similar enhancer was discovered between the miR-1-1 and mIR-133a-2 genes, indicating that MEF2C controls the expression of both bicistronic miRNA clusters via analogous regulatory elements [78]. Similarly, in Drosophila, an SRF-like enhancer element is required for miR-1 expression in cardiac progenitor cells [53], while Mef2 cooperates with the transcription factor Twist to regulate miR-1 expression in late-stage embryos [54]. Moreover, a genetic screen in flies revealed that miR-1 downregulates Cdc42 downstream of Tinman, the invertebrate ortholog of Nkx2.5; disruptions in this interaction severely impair adult cardiac physiology [80]. This interaction is also critical for maintaining cardiac contraction, electrical conduction, and rhythmicity in postnatal mice [80]. Taken together, the existence of highly conserved transcriptional networks directing miR-1/133a expression in the heart underscores the importance of these miRNAs in the regulation of cardiac development, as well as the maintenance of postnatal cardiac function. miR-1 and miR-133a are two closely related, striated muscle-enriched miRNAs that are transcribed together as members of a bicistronic cluster. The vertebrate genome contains two copies of the miR-1/133a locus on different chromosomes. The miR-1-1/133a-2 cluster is located within an intergenic region of chromosome 2, while the miR-1-2/133a-1 cluster is encoded within an intron of the protein-coding Mindbomb (Mib) gene. An additional intergenic miRNA cluster consisting of miR-206 and miR-133b is located on chromosome 1. This cluster is homologous to miR-1, but is expressed exclusively in skeletal muscle.
Loss-and gain-of-function studies in Drosophila were the first to demonstrate the specific role of miR-1 in the regulation of striated muscle development. Homozygous deletion of the single Drosophila miR-1 ortholog (dmiR-1) was lethal, with 100% of flies dying in larval stages [53,54]. Mutant flies were profoundly lethargic as second instar larvae, and exhibited complete collapse of the body wall musculature [54]. In a similar study, down-regulation of critical sarcomere genes was observed in half of dmiR-1 mutants, and severely affected embryos manifested defects in striated muscle patterning due to a failure of cardiac and skeletal muscle progenitors to terminally differentiate [53]. Overexpression of dmiR-1 in the late mesoderm disrupted the formation of cardioblasts, indicating that an excess of dmiR-1 may result in the premature differentiation of cardiac precursors, or divert cardiac progenitors towards an alternative cell fate [53]. Moreover, overexpression of dmiR-1 in the developing Drosophila wing disks markedly reduced the protein levels of the Notch ligand, Delta, suggesting that the muscle patterning defects observed in these studies arose from deregulated Notch signaling [53].
Similar studies have also uncovered a role for miR-1 in the regulation of cardiac development in various other model organisms. In mice, overexpression of miR-1 in cardiomyocytes arrested embryonic development by E13.5 secondary to heart failure and ventricular abnormalities [76]. In transgenic embryos, a significant reduction in the number of mitotic cardiomyocytes was observed relative to controls. This phenotype was consistent with premature cardiac progenitor differentiation, early withdrawal of cardiomyocytes from the cell cycle, and an insufficient expansion of ventricular cardiomyocytes [76]. Hand2, a cardiogenic transcription factor essential for cardiac differentiation, was further implicated as an in vivo target of miR-1. This finding supported a model in which miR-1 regulates Hand2 temporally and spatially during cardiac development [76].
Genetic knockout of miR-1-2 in mice provided the first evidence regarding the essential role of an individual miR-1 family member in cardiogenesis. Targeted miR-1-2 deletion on a pure Sv129 background resulted in incompletely penetrant lethality from E15.5 to birth, secondary to VSD [44,52]. The remaining homozygotes that survived to adulthood exhibited a spectrum of cardiac defects, including rapid cardiac dilation, atrial thrombosis, and various cardiac conduction abnormalities. Consistent with previous reports implicating miR-1 in the control of embryonic cardiomyocyte expansion [76], miR-1-2 null mice that survived to adulthood exhibited a 20% increase in the number of cardiomyocytes relative to wild-type controls. Remarkably, a significant increase in the number of cardiomyocytes undergoing nuclear division was also detected in mutant adults, a phenomenon not observed in wild type animals of the same age; concomitant increases in cardiomyocyte apoptosis were not detected [44]. Similarly, miR-1-1-null animals exhibited a number of cardiac defects, including conduction abnormalities and cardiac dilation, as well as incompletely penetrant lethality on a pure Sv129 background [52]. In contrast to these findings, however, a recent study found that targeted deletion of miR-1-1 or miR-1-2 caused neither embryonic lethality nor septal defects in Sv129 mice [55]. These phenotypic discrepancies may be attributed to residual pGK-neomycin and lacZ-pGK-neomyocin cassette expression in the previously reported miR-1-1 [52] and miR-1-2 [44] knockout mice, respectively [55].
Similar to the miR-1-1 and miR-1-2 single knockout mice described by Wei et al. [55], animals lacking either miR-133a-1 or miR-133a-2 in which the neomycin selection cassette had been excised showed no apparent defects in cardiac physiology, and responded to cardiac pressure overload in a manner indistinguishable from wild-type controls [56]. However, 50% of miR-133a-1; miR-133a-2 double knockouts died between P0 and P1 from severe VSDs, and only 25% of compound homozygotes survived to adulthood [56]. Consistent with severe defects in cardiac contraction and function, adult double knockout mice exhibited increased cardiomyocyte proliferation, enhanced expression of cell cycle-related genes, and ectopic expression of smooth muscle genes in the heart [56]. Moreover, cardiac-specific overexpression of miR-133a caused a spectrum of abnormalities in the developing heart, including diminished cardiomyocyte proliferation within the ventricular myocardium, and death by E15.5 secondary to cardiac failure. Many of these phenotypes were attributed to the upregulation of the genes encoding SRF and cyclin D2, which are both normally targeted for repression by miR-133a [56]. Taken together, these data indicate that miR-133a plays a critical role in the regulation of contractile gene expression and cardiomyocyte proliferation during embryogenesis, and suggest a possible role for miR-133a in mediating cardiomyocyte cell cycle withdrawal in early postnatal life.
Due to the fact that miR-1 and miR-133a possess different seed sequences, it is reasonable to assume that they regulate a separate set of target mRNAs and have distinct biological functions. Remarkably, however, there exists a significant functional overlap between the members of these two miRNA families with regard to their respective roles in cardiac development. For example, recent affinity purification data demonstrated that Hand2, a well-characterized target of miR-1, is also regulated by miR-133a in vitro and in vivo [81]. This study was important because it empirically established Hand2 as a functional miR-133a target, despite the inability of several miRNA prediction algorithms to identify this interaction [81]. Similarly, a study using zebrafish depleted of muscle-specific miRNAs showed that approximately 68% of upregulated transcripts possessed canonical 6mer, 7mer, or 8mer binding sites for miR-1 or miR-133a [57]. Moreover, concomitant morpholino-mediated knockdown of miR-1 and miR-133a manifested more severe disruptions in sarcomeric actin organization than knockdown of either miRNA alone, suggesting that miR-1 and miR-133a cooperatively regulate actin polymerization and sarcomere assembly [57]. Finally, in mice, genetic deletion of both miR-1/miR-133a clusters resulted in death by E11.5 secondary to severe impairments in blood circulation and cardiac function [58]. Consistent with reports of increased smooth muscle gene expression in miR-1-1/miR-1-2 [52] and miR-133a-1/133a-2 [56] double knockout hearts, transcriptional profiling of miR-1/133a compound knockouts revealed an upregulation of myocardin, which is a major regulator of smooth muscle gene expression [58]. Accordingly, although early stages of heart development such as linear heart tube formation proceeded normally in miR-1/133a mutant embryos, double knockout hearts consistently displayed a gene expression pattern characteristic of immature cardiomyocytes, in that they expressed both cardiac-and smooth-muscle specific transcripts [58]. Taken together, these findings indicate a novel role for the miR-1/133a bicistronic clusters in the transition of immature cardiomyocytes to a more mature phenotype, and illustrate the complexity of miRNA-mediated spatial and temporal modulation of gene expression.

miR-138 and miR-218 Control Cardiac Patterning
The acquisition of cardiac cell identity and the establishment of domain-specific function are dependent upon a complex network of transcriptional and post-transcriptional regulatory factors [82]. In vertebrates, heart development culminates in the formation of septated atrial and ventricular chambers, which exhibit unique gene expression patterns, and thus, possess distinct morphological, electrical, and functional properties. These chambers are separated by the atrioventricular canal (AVC), a discrete structure that gives rise to the cardiac valves and ultimately ensures the unidirectional flow of blood through the heart [83]. Studies using zebrafish have identified distinct gene expression profiles that distinguish the AVC from the surrounding cardiac chambers [84], and have uncovered a role for miRNAs in the regulation of associated cardiac patterning events. Specifically, when injected with a morpholino targeting miR-138, zebrafish embryonic ventricular cardiomyocytes failed to elongate, and instead, remained in a rounded, immature state reminiscent of cardiomyocytes within the AVC [59]. Moreover, genes normally confined to the AVC, such as notch1b and espg2, were ectopically expressed in the ventricle, indicating that miR-138 restricts the valve-forming region by repressing AVC-specific genes in maturing ventricular cardiomyocytes. These effects were attributed in part to miR-138-mediated downregulation of aldh1a2, which is required for the synthesis of retinoic acid, as well as the upregulation of espg2 in the AVC [59]. Although these findings have important implications for understanding the regulation of cardiac patterning in zebrafish, future studies are required to assess the requirement of miR-138 for the specification of cardiac domains in mammals.
Further support attesting to the importance of miRNAs in the regulation of cardiac morphogenesis and patterning has originated from studies of the miR-218 family in zebrafish. In mammals, the miR-218 family consists of two highly conserved members. miR-218a-1 and miR-218-a-2 are located within introns of the slit3 and slit2 genes, respectively; in fish, a third family member, miR-218b, is encoded intergenically. Although better known as an emerging cancer biomarker [85][86][87] and for its role in directing axon guidance during nervous system development [88,89], miR-218-mediated modulation of Slit-Robo (Roundabout) signaling is also critical for the patterning of the developing heart [60] and vasculature [90]. Injection of one-and two-cell stage zebrafish embryos with miR-218-targeting morpholinos impaired the migration of heart field progenitor cells to the midline during heart tube formation, and caused severe pericardial edema and cardiac morphological defects 48 hours post-fertilization (h.p.f.) [60]. Partial knockdown of robo1 rescued the morphant phenotype, implying that miR-218 regulates cardiac progenitor migration via direct robo1 repression [60]. However, using the same morpholinos as Fish et al. [60], Chiavacci et al. did not observe any cardiac cell migratory or morphological defects at 48 h.p.f., which was attributed to the extremely low levels of endogenous miR-218 expression at early developmental stages [61]. Accordingly, the authors reported abnormal cardiac looping, chamber malformations, and ectopic expression of tie2, a marker restricted to endothelial and hematopoietic cells, following miR-218 overexpression. Moreover, it was shown that miR-218 and tbx5 expression correlate in vitro and in vivo, and that many of the cardiac defects associated with tbx5 overexpression could be rescued by miR-218 downregulation [61]. Due to the fact that Tbx5 haploinsufficiency underlies the pathogenesis of Holt-Oram syndrome in humans, a congenital disease characterized by cardiac septation defects [91], further studies are required to assess the function of the miR-218 -Tbx5 regulatory circuit in vertebrate heart morphogenesis and patterning.

The miR-15 Family Negatively Regulates Cell Proliferation and Induces Embryonic Cardiomyocyte Mitotic Arrest
Cardiomyocytes undergo extensive proliferation during mammalian embryogenesis, yet withdraw from the cell cycle in early post-natal life [92]. This developmental switch is characterized by a transition from hyperplastic to hypertrophic cardiac growth, and is marked by the formation of binucleated cardiomyocytes, the maturation of intercalated discs, and an increase in myofibril density [93]. In order to identify miRNAs associated with this transition, Porrello et al. examined the developmental expression pattern of miRNAs in rodent ventricular cardiomyocytes as they approached cell cycle arrest [63]. These profiling experiments revealed a dramatic upregulation of multiple miR-15 family members, which corresponded with the onset of cardiomyocyte binucleation. Overexpression of the most highly upregulated miR-15 family member, miR-195, in the embryonic heart significantly reduced cardiomyocyte proliferation, and led to VSDs and ventricular hypoplasia in approximately 25% of neonatal mice. Moreover, miR-195 transgenic animals that survived to adulthood died prematurely due to the development of a slow-onset dilative cardiomyopathy [63]. Taken together, these data indicate that the miR-15 family plays an important role in the mediation of cardiomyocyte senescence in early post-natal life.
The propensity of miR-15 family members to negatively regulate cell proliferation has been attributed to their in vitro ability to directly target components of the core cell cycle machinery, such as cyclin D1, cyclin D3, and cyclin E [94,95]. In addition, experiments using a variety of cancer cell lines have revealed that the miR-15 and miR-16 clusters are direct targets of E2F, a critical transcription factor that mediates the transition from G1-to S-phase [95]. Interestingly, E2F is a downstream mediator of the tumor suppressor protein pRb, which was recently shown to maintain adult cardiomyocytes in a post-mitotic state by directing heterochromatin formation at the promoters of cell cycle-regulated genes [96]. Analysis of neonatal miR-195 transgenic mouse hearts using next-generation massively-parallel sequencing of RISC-associated transcripts (RISC-seq) further demonstrated that the miR-15 family targets a number of cell cycle-related genes in vivo [63]. One of these genes, Chek1 (checkpoint kinase 1), plays a number of roles during DNA repair and mitosis, including the prevention of genomic instability, and the coordination of G2/M progression [63]. Although this experiment provides some insight into the mechanisms by which the miR-15 family may regulate neonatal cardiomyocyte mitotic arrest, the function of Chek1 in cardiac development remains poorly understood, and therefore warrants further study.
Several studies have revealed important roles for the miR-17~92 cluster and its paralogs in cardiac development. Although mice homozygous null for either the miR-106b~25 or miR-106a~363 cluster did not display an overt phenotype, animals lacking the miR-17~92 locus died within minutes after birth due to VSDs and lung hypoplasia [66]. Moreover, compound deletion of miR-17~92 and miR-106b~25 was embryonic lethal, and resulted in a range of severe cardiovascular abnormalities including ventricular hypoplasia, atrial and ventricular septal defects, vascular congestion, and edema. The effects of miR-17~92 deletion were attributed in part to upregulation of pro-apoptotic proteins such as Bim, which was shown to be a direct target of this miRNA family [66]. Although further studies are required to more thoroughly assess the role of Bim deregulation on the cardiac phenotypes caused by miR-17~92 deletion, these results illustrate the essential and cooperative functions of miR-17~92 and its paralogs in cardiac development.
The miR-17-92 cluster has also been shown to influence the differentiation of second heart field progenitors, which contribute to the development of the right ventricle and outflow tract. During embryogenesis, cells of the second heart field are distinguished by the expression of the critical cardiogenic transcription factors Isl1 (ISL LIM homeobox 1) and Tbx1 (T-box 1) [98], which are transiently expressed prior to the initiation of cardiac progenitor cell differentiation [99,100]. In mice, compound deletion of Bmp2/4 (bone morphogenic protein) resulted in defective silencing of these cardiac progenitor genes, as well as a downregulation of cardiac differentiation markers [101]. This result was attributed to direct BMP-mediated regulation of pri-miR-17~92 expression, suggesting that BMP signaling is critical for the modulation of cardiac progenitor gene dosage [101].
Additional studies in zebrafish have revealed that miR-17~92 is an important regulator of cardiac morphogenesis and patterning. Specifically, injection of miR-92 mimics into the one-stage embryos was associated with reduced endoderm formation and cardia bifida, while morpholino-mediated miR-92 depletion resulted in left-right cardiac asymmetry defects [62]. Moreover, members of the miR-17~92 cluster have also been implicated in the regulation of organ size. Mice globally overexpressing a miR-17 transgene were smaller than their control littermates, and exhibited significant reductions in heart, liver, and spleen weight, likely due to miR-17-mediated repression of fibronectin expression [64]. However, overexpression of the entire miR-17~92 cluster in developing murine cardiomyocytes resulted in cardiac hyperplasia and hypertrophy, and caused sudden death by approximately two months of age [65]. The discrepancies between these two studies [64,65] suggest that members of the miR-17~92 cluster possess distinct functions, and illustrate the need for further studies to identify the roles of each individual miRNA in the regulation of cardiac development and function.

The myomiR Mediates Myosin Heavy Chain Isoform Switching during Fetal and Adult Stages and Helps Maintain Proper Cardiac Electrophysiology
Myosin is a major molecular component of cardiac muscle, and is the protein responsible for the generation of contractile force in that tissue. Cardiac contraction is primarily dependent upon the expression of two MHC (myosin heavy chain) genes, α and β, which are regulated in an antithetical manner in response to developmental and pathological signals [102]. In rodents, βMHC/Myh7, a slow adenosine triphosphatase (ATPase), is the predominant myosin isoform expressed in the embryonic heart, whereas αMHC/Myh6, a relatively faster ATPase, is highly upregulated during the early post-natal period [103,104]. By contrast, in humans and other large mammals, the expression of the βMHC isoform is retained during adulthood [102], suggesting that species-specific regulatory mechanisms underlie the control of myosin expression.
A recent series of findings has demonstrated that myosin isoform expression is regulated in part by a vertebrate-specific family of miRNAs known as the "myomiR." This miRNA family consists of miR-208a, miR-208b, and miR-499, which are encoded within the introns of Myh6, Myh7, and the closely related Myh7b, respectively. Interestingly, these miRNAs are expressed concomitantly with their host genes: In mice, miR-208b and miR-499 are most highly expressed during embryonic development, while miR-208a is upregulated shortly after birth [67,68,105]. Moreover, the members of the myomiR family possess closely overlapping "seed" sequences, indicating that they likely regulate a similar array of target transcripts.
Genetic loss-of-function studies in mice have demonstrated that individual myomiRs are not explicitly required for cardiac development. Animals null for miR-208a, miR-208b, or miR-499 are obtained at Mendelian ratios, and exhibit no overt developmental abnormalities or cardiac contractility defects prior to weaning [67,68,105]. However, the high degree of sequence similarity among these three miRNAs may confer functional redundancy, and further experiments employing compound knockout animals are required to fully assess the role of the myomiR trio in cardiogenesis. Interestingly, however, one study reported a potential role for miR-208a in the maintenance of cardiac contraction during adult life. Specifically, miR-208a-deficient animals experienced progressive declines in cardiac function beginning at approximately two months of age, and manifested severe abnormalities in sarcomere structure and ectopic expression of skeletal muscle-specific genes at six months of age [67]. Consistent with defects in cardiac contraction, electrocardiography (ECG) readings from four month-old miR-208a knockout mice lacked P-waves [68]. Given that P-waves normally represent atrial depolarization and contraction, the loss of P-waves in miR-208a knockout mice is suggestive of atrial fibrillation [68]. Likewise, cardiac-specific overexpression of miR-208a resulted in an elongation of the PR intervals in transgenic mice, indicative of what is clinically considered first-degree AV (atrioventricular) block. These electrophysiological abnormalities were attributed to interactions between miR-208a and the gap junction protein Cx40 (connexin 40), and transcription factors normally expressed in the cardiac conduction system such as GATA4 (GATA binding protein 4) and Hop (homeodomain-only protein) [68]. However, contrary to the findings of van Rooij et al. [67] and Callis et al. [68], another study demonstrated that systemic administration of miR-208a antimiRs in adult mice did not induce changes in cardiac conduction properties [106]. Therefore, further experiments are required to more clearly define the role of miR-208a and the remaining members of the myomiR family in the maintenance of proper cardiac electrophysiology.
Although loss of miR-208a manifests only minor defects with regard to cardiac development and function, this miRNA plays a crucial role in mediating certain aspects of the cardiac stress response. Reactivation of the fetal cardiac genetic program, including myosin isoform switching, is a characteristic of the stressed, hypertrophic, or failing post-natal heart [107]. In response to TAB (thoracic aortic banding), a procedure that reliably induces hypertrophic growth in the hearts of wild-type mice by increasing cardiac afterload, miR-208a-deficient animals displayed minimal signs of cardiac hypertrophy or fibrosis [67]. Although the expression of cardiac stress markers, such as atrial natriuretic factor (ANF) and b-type natriuretic peptide (BNP), was unaffected in the hearts of miR-208a knockout mice, these animals were remarkably unable to upregulate βMHC [67,68]. Similarly, cardiomyocyte-specific overexpression of miR-208a was sufficient to induce βMHC expression in the heart, even in the absence of overt cardiac pathology. These effects were proposed to result in part from miR-208-mediated repression of THRAP1 (thyroid hormone receptor-associated protein 1), a transcriptional co-regulator of thyroid hormone signaling [67,68], and myostatin, a suppressor of skeletal muscle hypertrophy [68]. Due to the fact that loss of miR-208a during embryogenesis has no effect on βMHC protein levels [67,68,105], it will be interesting to determine if distinct regulatory mechanisms underlie βMHC expression during development and situations of cardiac stress.

MiRNAs and Proliferation Endogenous Cardiomyocytes
Historically, the heart was considered a terminally differentiated organ incapable of proliferation or regeneration. However, this dogma has been challenged by recent reports of limited yet significant cardiomyocyte turnover in adults during aging [108,109], as well as following myocardial infarction [109]. Nevertheless, the intrinsic regenerative capacity of the adult mammalian myocardium is not sufficient to compensate for the loss of cardiac function post-injury, and ischemic cardiac disease remains the leading cause of adult morbidity and mortality worldwide. In contrast to the negligible innate regenerative ability of the adult heart, Porrello et al. recently demonstrated that the hearts of neonatal mice possess a robust capacity for regeneration during the first week of life in response to resection of the ventricular apex [92] and myocardial infarction [110]. Interestingly, this response is lost in seven day-old mice, a time point corresponding to the withdrawal of cardiomyocytes from the cell cycle [92,110].
The proliferation of existing cardiomyocytes has been identified as the primary mechanism underlying cardiac regeneration in neonatal mice [92,110] and lower vertebrates such as zebrafish [111] and amphibians [112]. Due to the important role of the miR-15 family in the modulation of cardiomyocyte proliferation during cardiogenesis, the functions of these miRNAs have also been investigated in the setting of cardiac regeneration. Neonatal transgenic animals overexpressing the miR-15 family member miR-195 in cardiomyocytes fail to mount a regenerative response to myocardial infarction, and exhibit features characteristic of adult cardiac remodeling including inflammation, cardiomyocyte hypertrophy, and fibrosis [110]. Similarly, systemic antimiR-mediated inhibition of the miR-15 family from birth to adulthood induced cardiomyocyte proliferation in infarcted adult hearts, and elicited improvements in left ventricular systolic function [110]. Although these data indicate that members of the miR-15 family may be valid targets for patients with ischemic cardiac disease, systemic antimiR administration may elicit off-target effects. Therefore, further studies that evaluate the safety and efficacy of cardiomyocyte-specific miR-15 inhibition are required before this approach is attempted in clinical trials.
The observation that the miR-17~92 cluster is required for the proliferation of cardiomyocytes in embryonic, postnatal, and adult hearts [113] suggests that these miRNAs may be attractive therapeutic targets for patients with cardiac disease. To this end, transgenic embryos overexpressing miR-17~92 in the heart displayed enhanced cardiomyocyte proliferation relative to control littermates. Moreover, overexpression of the miR-17~92 cluster induced adult cardiomyocyte proliferation in healthy animals, as well as in response to myocardial infarction [113]. These effects were attributed to miR-17~92-mediated inhibition of the tumor-suppressor protein PTEN (phosphatase and tensin homolog) [113]. However, due to the role of this miRNA cluster in tissue growth and tumorigenesis, further studies evaluating the safety of cardiomyocyte-specific miR-17~92 overexpression are required before these data can be applied to human patients.
Additional support for the role of miRNAs in cardiac regeneration has originated from a functional screen seeking to identify mature miRNAs capable of inducing cardiomyocyte proliferation [114].
Forty miRNAs were shown to enhance DNA synthesis in neonatal rodent cardiomyocytes, while two miRNAs, miR-199a and miR-590, were capable of inducing cell-cycle re-entry in adult rat cardiomyocytes in vitro. Moreover, AAV9 (adeno-associated virus serotype 9)-mediated in vivo delivery of miR-199a or miR-590 elicited a significant regenerative response in adult animals following myocardial infarction [114]. This study provides compelling evidence for the utility of miRNAs to in vivo cardiac regeneration, and it will be interesting to evaluate the roles of endogenous miR-199a and miR-590 during cardiac development.

MiRNAs and Cell-Based Strategies for Cardiac Regeneration
The field of stem cell biology is rapidly expanding and holds great potential for cardiac regeneration. Indeed, cardiomyocytes derived from embryonic stem cells (ESCs) and resident cardiac progenitor cells (CPCs) have imparted considerable functional improvements in animal models of ischemic heart disease [115]. However, clinical trials of these cell-based therapies have only reported modest, transient results, and the benefits of stem cells for cardiac regeneration in human patients with ischemic heart disease have not been conclusively established. Further studies are required to improve the safety of cell transplantation, and to overcome challenges such as poor engraftment rates, immunomodulation of engrafted cells, and patient heterogeneity [116]. One approach to improving the feasibility of these therapies for heart disease has been to incorporate miRNAs involved in the regulation of stem cell potency and self-renewal, as well as cardiomyocyte lineage commitment and differentiation. Although numerous studies attest to the potential of miRNAs to enhance cell transplantation therapies for cardiac disease, these findings have been reviewed extensively elsewhere [117][118][119][120], and will not be a focus of this review.
An alternative strategy to the use of ESCs and CPCs for cardiac regeneration is lineage reprogramming, an approach whereby one cell type is effectively converted into another. Historically, it was believed that the differentiation of somatic cells represented a permanent restriction of cell plasticity [121]. However, this dogma was overturned by somatic cell nuclear transfer experiments, which demonstrated that the nucleus of a somatic cell could revert to a pluripotent state and support the development of an entire embryo when transplanted into an enucleated oocyte [122,123]. An explosion in the cell reprogramming field occurred following Shinya Yamanaka's landmark discovery that embryonic and adult murine fibroblasts could be induced to pluripotency by the introduction of the transcription factors Oct3/4, Sox2, c-Myc, and Klf4 [124]. Interestingly, expression of the miR302/367 cluster in the presence of the HDAC inhibitor valproic acid was shown to revert mouse and human fibroblasts to pluripotency more efficiently than the original reprogramming factors [125]. These induced pluripotent stem cells (iPSCs), which can be differentiated into virtually any somatic cell type following the addition of lineage-specific signaling molecules, have revolutionized the field of regenerative medicine. They represent a source of patient-specific stem cells amenable to disease modeling, drug discovery, and the development of personalized therapies [126].
In attempt to address concerns regarding the tumorigenic potential of iPSCs and the immaturity of iPSC-derived cardiomyocytes, several groups have adopted a direct reprogramming strategy, in which one somatic cell type is transdifferentiated into another. The mammalian heart is a rich reservoir of cardiac fibroblasts, which function as a structural scaffold for cardiomyocytes during embryogenesis [127], and contribute to extracellular matrix homeostasis in adults [128]. Cardiac fibroblasts also mediate the response to ischemic cardiac injury by proliferating and migrating to the infarct area where they orchestrate tissue remodeling, collagen deposition, and scar formation [129]. Ultimately, the loss of contractility within the infarcted region contributes to ventricular dysfunction, mechanical cardiac overload, and the development of chronic heart failure [130,131]. Due to their relative abundance in the heart and their fundamental role in the response to cardiac injury, Ieda et al. speculated that cardiac fibroblasts could represent a source of cardiomyocytes for regenerative therapies [132]. They systematically screened a group of key cardiogenesis regulators in order to identify factors capable of directly reprogramming cardiac fibroblasts into cardiomyocytes; ectopic expression of the transcription factors Gata4, Mef2C, and Tbx5 (GMT) was necessary and sufficient to generate cardiomyocyte-like cells in vitro [132]. Although the generation of functional, beating cells occurred at a very low frequency, Qian et al. [133] and Song et al. [134] subsequently showed that GMT-mediated reprogramming could also occur in vivo, and impart significant functional improvements in murine models of myocardial infarction.
Despite the apparent promise of lineage reprogramming to cardiac regeneration, many of these approaches rely on the use of virally encoded transcription factors. This feature of current reprogramming technology raises concerns due to the possibility of insertional mutagenesis [135]. In order to improve the safety of reprogrammed cells, alternative reprogramming strategies that circumvent the use of integrating viral vectors are being developed. For example, it was recently shown that when transiently transfected with miR-1, -133, -208, and -499, cardiac fibroblasts could acquire a gene and protein expression profile consistent with that of cardiomyocytes [136]. This study also provided proof-of-principle evidence that this miRNA cocktail could mediate direct cardiac reprogramming in a mouse model of cardiac injury [136]. However, because this latter experiment employed lentivirally-encoded miRNAs, it is not clear whether in vivo reprogramming can be accomplished with transiently transfected cardiac fibroblasts. Moreover, the authors did not indicate whether the induced cells were capable of improving cardiac function after injury. Therefore, additional functional data are required to fully assess the utility of miRNAs to in vivo direct cardiac reprogramming.

Synthesis and Prospects
Since the initial discovery implicating miRNAs as essential components of the cardiogenic regulatory network, numerous roles for these small RNAs in cardiac development and disease have been identified. However, of the over 800 miRNAs annotated to-date, only a minute fraction has been studied in the context of the heart. Current bioinformatics algorithms have contributed immensely to our understanding of miRNA function. However, the inability of these models to accurately predict miRNA targets represents a significant challenge that must be addressed. A greater understanding of miRNA mechanisms of action, including the role of non-seed regions, will improve our ability to identify miRNA-mRNA interactions. Moreover, the recent discovery that miRNAs can also function as transcriptional activators may reveal further layers of complexity with respect to the relationships between miRNAs and their targets.
Although we now recognize that miRNAs function in virtually every facet of cell physiology, the field of miRNA biology is only just emerging. Undoubtedly, the expansion of this area of research will uncover novel prospective therapeutic targets for individuals with cardiovascular disease. Further studies seeking to enhance our understanding of the role of miRNAs in cardiogenesis and cardiac regeneration, as well as those designed to ameliorate current experimental and therapeutic methodologies, are required to develop new therapies that impart significant clinical benefits to patients with congenital and acquired cardiac disorders.