REVIEW URRENT C OPINION
Therapeutic applications of noncoding RNAs Michael V.G. Latronico a and Gianluigi Condorelli a,b,c
Purpose of review In this review, we summarize the basic principles underlying the therapeutic use of nonprotein coding (nc)RNAs, such as microRNA (miRNA) and long noncoding RNA, in the cardiovascular field, focusing, where possible, on recent advances that may lead to translation to the clinic for heart failure. Recent findings The number of individual miRNAs associated with a given aspect of heart disease is increasing rapidly, as is the data on miRNA profiles in normal and diseased myocardium. Less is known on the role of long noncoding RNA, and to date only a few have been studied in the heart. Novel oligonucleotide-based therapies have started to trickle into the clinic, but strategies focusing on ncRNA are still in a clinical/ preclinical trial phase. Summary The discovery of functional ncRNAs is leading to a better understanding of the mechanisms underlying cardiovascular physiology. Dysregulation of ncRNAs is being increasingly associated with many diseases affecting the heart and in certain instances may have a pathogenic role. Therapeutic intervention aimed at opposing ncRNA misexpression has been widely demonstrated to be effective in blunting disease in animal models, and may thus have potential in the clinical setting. Keywords heart, long noncoding RNAs, microRNAs
INTRODUCTION Here, we give an outline of the progress made in the use of nonprotein coding RNA (ncRNA) (see supplementary text and figure, http://links.lww.com/ HCO/A25) for therapeutic ends, with a focus on heart failure where possible. We will first review how the dysregulation of microRNAs (miRNAs) associated with the development of disease may be treated, and then go on to briefly discuss long noncoding RNA (lncRNA).
THERAPEUTIC MODULATION OF miRNA ACTIVITY Many studies have documented the potential of miRNAs in the field of cardiovascular therapeutics [1–4]. The advantages afforded by miRNA-based therapeutics include the fact that miRNA sequences are short and often highly conserved, making them relatively easy not only to target and manipulate but also to study in the various preclinical animal models adopted. They may also modulate a number of players in a given pathway, an effect that potentiates any therapeutic effect on a disease pathway; this is an important point because the effect of a given miRNA on a single mRNA target may be only
slight [5]. Moreover, whereas miRNAs themselves are particularly stable for an RNA (half-life ranging from 24 h to 5 days) even outside the cell [6], miRNA-based therapeutics may have relatively long-lasting effects, especially if designed with improved pharmacokinetic and pharmacodynamic properties. Depending on whether the disease is associated with an increased or a decreased expression of the miRNA, two general approaches can be employed to modulate the activity of the ncRNA: miRNA mimics may be used to restore the function of disease-downregulated miRNAs, whereas any increase in the expression of an miRNA can be inhibited with complementary oligonucleotide sequences.
a
Humanitas Research Hospital, bUniversity of Milan and Research Council of Italy, Rozzano, Milan, Italy
c
National
Correspondence to Gianluigi Condorelli, MD, PhD, Cardiovascular Research Laboratories, Humanitas Clinical and Research Center, via Manzoni 56, 20089 Rozzano, Milan, Italy. Tel: +34 02 8224521; fax: +34 02 82245404; e-mail:
[email protected] Curr Opin Cardiol 2015, 30:213–221 DOI:10.1097/HCO.0000000000000162
0268-4705 Copyright ß 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.co-cardiology.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Molecular genetics
KEY POINTS Functional ncRNAs – such as miRNAs and lncRNAs – make up a substantial fraction of the transcriptome and generate a complex regulatory network, the individual parts of which may each potentially induce a pathological state when misexpressed either positively or negatively. A disease associated with pathogenic downregulation of a particular ncRNA can in theory be treated via the use of ncRNA mimics to recover its physiological level; similarly, disease-upregulated ncRNAs may be targeted with ASOs that sequester them, inhibiting their functioning. The first antisense drugs targeting mRNAs are making a timid entry into the clinic, also for cardiovascular diseases, but strategies focusing on miRNAs are still in the preclinical phase. Studies on the targeting/use of lncRNAs for therapeutic ends in the cardiovascular field are less advanced.
RESTORING miRNA FUNCTION Replacement therapies for a disease-downregulated miRNA can be implemented via overexpression of the miRNA with the use of viral vectors – which introduce gene copies of the therapeutic miRNA into the host – or via the use of synthetic doublestranded oligonucleotides that resemble miRNA duplexes and so can directly co-opt the miRNAmediated silencing complex (miRISC) machinery (Fig. 1).
miRNA mimics Although double-stranded miRNA mimics are now relatively easy to synthesize, the polyanionic nature of oligonucleotides makes them difficult to pass across the cytoplasmic membrane of cells targeted for therapy; moreover, RNAs are particularly susceptible to enzymatic degradation in the extracellular and intracellular milieu. To overcome these limitations, this type of miRNA-based medicinal must be chemically modified with respect to the native miRNA molecule (Fig. 2). The modifications available to date include conjugation with cholesterol, which enhances cellular uptake (useful for the complete double-stranded mimic) [7], 20 -fluoro (20 -F) modifications, which increase stability in serum while not affecting loading into the RISC (useful for the guide strand) [8], and modifications that prevent RISC loading, ensuring rapid degradation (useful for the passenger strand) [9]. A recent demonstration of how reconstitution of miRNA function may abate the development of a 214
www.co-cardiology.com
pathologic state in vivo was given by the van Rooij laboratory with an investigation on miR-29 in pulmonary fibrosis [10]. miR-29 is a well studied miRNA family that regulates multiple extracellular matrix proteins (reviewed in [4]). Fibrosis is associated with downregulation of this miRNA, an event that derepressed collagen, elastin, and fibrillin gene expression. miR-29 is thus an example of an miRNA targeting several mRNAs associated with a specific disease pathway. Under basal conditions, intravenous administration of synthetic RNA duplexes was well tolerated by mice and was able to increase miR-29 levels robustly in various organs (including heart) already on the day after the injection. Importantly, the development of pulmonary fibrosis was blunted when the duplexes were administered after experimental induction of the disease via intratracheal instillation of bleomycin. The therapeutic effect was associated with significant reduction in bleomycin-stimulated increases in Col1a1, Col3a1, and Igf1. Thus, the authors were able to demonstrate the therapeutic applicability of using this strategy to restore the function of lost or downregulated miRNAs, at least in lung [10]. The miR-29b mimic used was constructed with chemical modifications for stability and improved cellular uptake. The guide strand was identical to native miR-29b, apart from a UU overhang to increase stability, a chemically phosphorylated 50 end, and 20 -F modification to help protect against exonucleases. However, because the guide strand is the functional strand, the modifications that could be implemented were limited so as not to interfere with its functionality. In contrast, the passenger strand was extensively modified with cholesterol conjugation (to increase cellular uptake), 20 -O-methylation (20 -O-Me) modifications (which prevent loading into the miRISC), and mismatches to impede its functioning as an miRNA. As an aside, artificial miRNAs may theoretically be used as a therapeutic option to downregulate one or more disease-upregulated proteins for which a regulating miRNA is not known; to this end, an online tool (called miR-Synth) for the design of artificial miRNAs based on user-selected targets and nontarget sequences has recently been presented by the Croce laboratory [11 ]. &
Viral vector-mediated strategy The use of recombinant viral vectors – in particular those based upon the nonpathogenic parvovirus adeno-associated single-stranded DNA virus (AAV) – has been demonstrated in animal models to be able to drive robust in-vivo miRNA expression characterized by long-lasting transgene expression Volume 30 Number 3 May 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Therapeutic applications of noncoding RNAs Latronico and Condorelli
Co-optation of miRNA biogenetic pathway
Transgene transcription by host cell
Dicer Duplex-type mimic
Cho l
Vector-delivered transgene
Disease-downregulated mIRNA miRNA Argonaute
Target mRNA
Recovery of lost translational downregulation
FIGURE 1. Strategies that may be used for the recovery of microRNAs (miRNAs) that become downregulated upon disease. Viral vectors may be used to deliver miRNA-expressing transgenes to the target cell/tissue in a relatively stable manner. The expressed trangene then co-opts the cell’s miRNA biogenesis pathway to recover lost miRNA expression levels. In addition, duplex-type mimics may be administered that can enter into cells [on account of modifications such as conjugation with cholesterol (Chol)] and co-opt the miRISC machinery.
and little immunogenicity [12–14]. AAV serotypes have been found with improved efficiency for the transduction of specific tissues: serotype 9 (and to a lesser extent AAV 8) seems particularly cardiotrophic when administered systemically [15], transducing cardiomyocytes and neurons – both postmitotic cell types – as well as skeletal muscle cells, while leaving many other proliferating cells unaffected; serotypes 1 and 6 may also be used if delivery is intramyocardial, intrapericardial, or intracoronary, which clinically may be more desirable routes on account of the fact that transduction of the myocardium can be obtained with the use of lower dosages [16]. Still, the bioengineering of novel AAV variants, such as via viral capsid mutagenesis,
may very well be able to produce AAV with an even more focused cardiac tropism [17]. Thus, with respect to the use of double-stranded oligonucleotides, the viral vector-mediated approach has greatly improved (cell/tissue) targetability and reduced problems with introduction of the medicinal into the cell. In addition, self-complementary AAVs have also been developed; these have increased efficiency in the heart, but utilize smaller, dimeric inverted repeats, and so are useful only for small transgenes, such as those of miRNAs [18]. The fact that they pack an already complementary double-stranded DNA means that second-strand synthesis is not necessary, so they lead to more rapid transgene expression.
0268-4705 Copyright ß 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.co-cardiology.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
215
Molecular genetics
HO
HO
Base
Base
O
O
RNA O
LNA
OH
O
O P O O
O
O P O O–
Base O
Base O
2´-Me
PMO
OCH3
O O P O O–
O
Base O
PO
2´-MOE OCH3
O
O O P O S–
O P N O–
Base Base
N
O
PS
2´-F
O
PNA
NH O
PO4
F
n
FIGURE 2. Nucleic acid variants used to modify the characteristics of oligonucleotides. The variations can involve modification of the sugar structure at the 20 position of RNA to form 20 -O-methyl (20 -O-Me), 20 -O-methoxyethyl (20 -O-MOE), and 20 -fluoro sugars (in light grey), or a locked nucleic acid (LNA); modification of the normal phosphodiester (PO) bond between sugars to form phosphorothioate (PS) bonds (in grey); or use of a nonribose backbone, such as with phosphorodiamidate morpholino oligomers (PMO) or peptide nucleic acids (PNA).
AAV vectors are relatively easy to design because of the presence of a small genome; only the inverted terminal repeats and the two viral genes rep (replication) and cap (capsid, which also harbors a third gene, AAP) of the wild-type virion (contained in an overall length of 4.7 kb) are necessary for replication and encapsidation. In the recombinant AAV vector, rep and cap are replaced by a promoter and the transgene; for virion production, these genes are provided in trans with a helper plasmid without inverted terminal repeats. The transgene cassette available for the production of recombinant AAV can accommodate a number of short-hairpin-sized genes or a single, longer transgene [19]. Transcription of a 60–70-bp-long transgene can be directed by a polymerase III-mediated mechanism to produce a transcript that can co-opt the miRNA biosynthetic machinery, generating small interfering RNAs; however, the high, ubiquitous activity of these promoters may produce saturation of the endogenous miRNA pathway and cannot provide tissue specificity [20]. More tissue specificity can be obtained with polymerase II-based promoters, such as a-myosin heavy chain, myosin light chain 2, and cardiac troponin C, albeit at the price of efficiency [21]. In addition, transgene silencing can be produced in specific cell types by the introduction of an miRNA target sequence in the transgene; for example, offtarget expression in liver and skeletal muscle can be 216
www.co-cardiology.com
reduced by the insertion of sites recognized, respectively, by miR-122 [22] or miR-206 [23 ], miRNAs expressed specifically in these districts. In contrast to the wild-type virion’s genes, a transgene loaded into a recombinant AAV does not integrate into the host’s genome but persists as double-stranded extrachromosomal DNA (in either linear or supercoiled circular form) almost indefinitely in nondividing cells [24]. This may lead to an extended therapeutic window, probably in the tens-of-years range for humans [25]. Examples of the use of viral vectors in the cardiovascular system include AAV9 expressing miR-590-3p and miR-199-3p, which have been shown to retrigger cardiomyocyte proliferation and regeneration after myocardial infarction [26], and those expressing miR-378 or miR-669a, which attenuate cardiac enlargement because of pressure overload [27] or cardiomyopathy [28], respectively, in mice. Another example of how the use of an AAV-mediated strategy can be employed to recoup miRNA function was demonstrated in rats 2 weeks after the induction of pressure overload cardiac hypertrophy [29]; rats administered an intravenous injection of an AAV9 vector harboring a hairpin precursor flanked by the native intron sequences of miR-1 – an miRNA that becomes downregulated in hypertrophic cardiomyocytes – underwent regression of cardiac hypertrophy and fibrosis, &
Volume 30 Number 3 May 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Therapeutic applications of noncoding RNAs Latronico and Condorelli
had less cardiomyocyte apoptosis, and showed improvements in calcium handling; of note, chronic restoration of miR-1 expression (up to 7 weeks) prevented deterioration of cardiac function in the face of elevated systolic pressure because of the continual presence of the aortic constriction that had generated the pressure overload.
binding to the natural mRNA target(s) so as to recover the physiological protein expression level(s) (Fig. 3). These antisense sequences must have a binding affinity sufficient to overcome the RISC’s helicase activity – which would otherwise remove it from the targeted miRNA – and be resistant to cleavage by argonaute (AGO). In fact, it has recently been shown in mice that anti-miRNA physically associate, via seed region pairing, with AGO-bound targeted miRNA, preventing association with the cognate mRNA(s) via a steric hindrance mode of action [31].
INHIBITING miRNA FUNCTION Downregulation of miRNA function has been more studied with respect to the gain-of-function approach [30]. Basically, the strategy is to increase the cellular pool of miRNA-recognizing sites, effectively ‘soaking up’ specific disease-upregulated miRNAs and thus reducing inappropriate miRISC
Antisense oligonucleotides Antisense oligonucleotides (ASOs) – a technology that has been used for more than 25 years to
AntagomiR Tiny LNA
AMO
Sponge
Transgene transcription by host cell
Vector-delivered transgene
Disease-upregulated miRNA Ch ol
Target mRNA
Sequestration of excess miRNA leading to blunting of inappropriate translational downregulation and recovery of normal protein level
FIGURE 3. Strategies that may be used to downregulate the expression of microRNAs (miRNAs) that become upregulated upon disease. Anti-miRNAs may be delivered either directly as chemically modified oligonucleotides [such as antagomiRs or tiny locked nucleic acids (LNAs)] or as a vector-delivered transgene. 0268-4705 Copyright ß 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.co-cardiology.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
217
Molecular genetics
experimentally blunt expression of mRNAs – can be employed against miRNAs. ASOs are synthesized to be complementary to the target, and so bind to it with high affinity, interfering with downstream mechanisms. The basic methodology is relatively easier to execute than gain-of-function approaches; miRNA inhibitory sequences – called anti-miRNA oligonucleotides (AMOs) or anti-miRNAs – need only be single-stranded, completely complementary oligonucleotides to the targeted miRNA, and can be chemically modified to improve cell entry and resistance to serum nucleases without worrying about the need to retain incorporability into the RISC, and thus obviating the need for potentially problematic delivery systems such as viral vectors. Many chemical modifications that can be used to improve anti-miRNA pharmacodynamics and pharmacokinetics are similar to those already described for the mimics (Fig. 2). Cholesterol conjugation to the 30 end of the sequence was found to improve uptake of so-called antagomiRs (asymmetrical, phosphorothioate- and 20 -O-Me-modified, fully complementary oligonucleotides to the cognate miRNA sequence) [7]; modifications at the 20 position of the sugar molecule include 20 -O-Me, 20 -O-methoxyethylation, and 20 -F modifications to confer increased resistance to nucleases [32–34]. The parent phosphodiester backbone linkages can also be modified to improve resistance to exonucleases present in serum, thus permitting systemic administration, by converting them into phosphorothioate [35] or replacing them with either a peptide-like backbone consisting of N-(2-aminoethyl)glycine units (peptide nucleic acids) [36] or phosphorodiamidate morpholino oligomers, in which a six-membered morpholine ring replaces the sugar moiety [35]. Modification of the phosphorothioate backbone can also enhance binding to plasma proteins, leading to reduced urinary clearance and thus improved pharmacokinetics [37]; however, backbone modifications may decrease the inhibitory action of the anti-miRNA, so they must be individually assessed and used selectively along the oligonucleotide. In addition, a 20 -O,40 -C methylene bridge can be used to lock the furanose ring of the sugar-phosphate backbone [producing locked nucleic acids — (LNAs)] to improve target affinity [38,39]. This last modification is so effective that LNAs can be synthesized with shorter sequences (15–16 nts) than normal anti-miRNAs. Tiny LNAs have also been designed to be complementary only to 7–8 nts of the seed sequence of the targeted miRNA, and can inhibit entire miRNA families [40]; in heart, these have been shown to inhibit miRNA families more effectively than via the use of several AMOs targeting the individual miRNAs. 218
www.co-cardiology.com
Moreover, in pigs, one intravenous injection of LNA was able to greatly knock down miRNA expression for several days throughout the heart as early as 1 day after administration [41]. A more recently developed type of modification involves nonnucleotide modifiers; for example, incorporation of naphthyl-based modifying groups, such as N,N-diethyl-4(4-nitronaphthalen-1-ylazo)-phenylamine (ZEN), increased binding affinity and blocked exonuclease degradation when placed at or near each end of a single-stranded 20 -O-Me-modified phosphodiester oligonucleotide [42]. However, their efficacy remains to be reported in vivo. A demonstration of the use of an anti-miRNA to inhibit an miRNA family in the hypertrophic heart was given by the McMullen laboratory [43]. An 8-nt-long LNA (with a complete phosphorothioate backbone) targeting the miR-34 family seed was delivered in repeated doses subcutaneously to mice with cardiac hypertrophy. Only administration of an LNA able to target the entire miR-34 family was capable of improving systolic function and a number of other functional, morphological, and molecular parameters in mice with preexisting cardiac dysfunction. No adverse effects were noted in other tissues.
Vector-mediated strategies Despite the possibilities produced by chemical modification, the therapeutic action of ASOs may be too transient, so repeated administrations would probably be necessary to maintain any prolonged effect. This is especially important for the heart failure setting, in which any inability to tackle the pathological alterations upstream of this syndrome would entail persistence of the prohypertrophy stimuli. As for the gain-of-function approach, a viral vector could also be used in this case. One possible method that could be used to increase the length that an anti-miRNA is expressed, and also augment its affinity, is the use of so-called ‘miRNA sponges’ [44]. Sponges are RNAs containing multiple (4–10) target sites fully/partly complementary to an miRNA of interest; they can be designed to neutralize a whole miRNA family, and are introduced into the cell as a transgene. These have been used for experimental purposes, but in theory they could be designed with regulatory elements in the promoter to make it drug-inducible or tissuespecific, and so become attractive for therapeutic ends. The first sponge-like anti-miRNA used to downregulate an miRNA in the heart of mice was the miR-133 decoy, an adenovirally mediated antimiRNA harboring two perfectly complementary binding sites downstream of a CMV promoter Volume 30 Number 3 May 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.