Journal of Controlled Release 172 (2013) 962–974

Contents lists available at ScienceDirect

Journal of Controlled Release

j ourna l homepage: www.e lsev ie r .com/ locate / jconre l

Review

Progress in microRNA delivery

Yu Zhang a, Zaijie Wang a, Richard A. Gemeinhart a,b,c,⁎a Department of Biopharmaceutical Sciences, University of Illinois, Chicago, IL 60612-7231, USAb Department of Bioengineering, University of Illinois, Chicago 60607-7052, USAc Department of Ophthalmology and Visual Sciences, University of Illinois, Chicago 60612-4319, USA

Abbreviations:AMO, anti-miRNAoligonucleotides; APdimethyldioctadecylammonium bromide; DGCR8, DiGeorFANA, fluorine derivatives nucleic acid; GC4, phage identiIR, ionizing radiation; LAC, lung adenocarcinoma; LNA, loclipid emulsion; NSCLC, non-small cell lung cancer; nt, nucacid; RISC, RNA induced silencing complex: scFv, single-cliposomes; TPGS, D-alpha-tocopheryl polyethylene glycol⁎ Corresponding author at: 833 South Wood Street (M

E-mail address: rag@uic.edu (R.A. Gemeinhart).

0168-3659/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.jconrel.2013.09.015

a b s t r a c t

a r t i c l e i n f o

Article history:Received 18 April 2013Accepted 15 September 2013Available online 25 September 2013

Keywords:MicroRNAmiRNA therapeuticsSmall RNA deliveryHuman diseaseNonviral delivery

MicroRNAs (miRNAs) are non-coding endogenous RNAs that direct post-transcriptional regulation of gene ex-pression by several mechanisms. Activity is primarily through binding to the 3′ untranslated regions (UTRs) ofmessenger RNAs (mRNA) resulting in degradation and translation repression. Unlike other small-RNAs, miRNAsdo not require perfect base pairing, and thus, can regulate a network of broad, yet specific, genes. Although wehave only just begun to gain insights into the full range of biologic functions of miRNA, their involvement inthe onset and progression of disease has generated significant interest for therapeutic development. Mountingevidence suggests that miRNA-based therapies, either restoring or repressing miRNAs expression and activity,hold great promise. However, despite the early promise and exciting potential, critical hurdles often involving de-livery of miRNA-targeting agents remain to be overcome before transition to clinical applications. Limitationsthat may be overcome by delivery include, but are not limited to, poor in vivo stability, inappropriatebiodistribution, disruption and saturation of endogenous RNA machinery, and untoward side effects. Both viralvectors and nonviral delivery systems can be developed to circumvent these challenges. Viral vectors are efficientdelivery agents but toxicity and immunogenicity limit their clinical usage. Herein, we review the recent advancesin the mechanisms and strategies of nonviral miRNA delivery systems and provide a perspective on the future ofmiRNA-based therapeutics.

© 2013 Elsevier B.V. All rights reserved.

Contents

1. Discovery and action of miRNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9632. MicroRNA therapeutic approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963

2.1. Chemical modification to RNA for miRNA therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9632.2. miRNA inhibition therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9652.3. MicroRNA replacement therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 966

3. Synthetic materials for miRNA and anti-miRNA oligonucleotide delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9663.1. Lipid-based delivery system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9663.2. Polyethylenimine (PEI)-based delivery system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9683.3. Dendrimers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9683.4. Poly(lactide-co-glycolide) (PLGA) particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9693.5. Other nonviral miRNA therapeutics delivery system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 969

4. Future outlook: challenges and opportunities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9705. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 971Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 971References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 972

P, amyloid precursor protein; BBB, blood–brain barrier; bp, base pairs; CNS, central nervous system; CSC, cancer stemcell; DDAB,ge syndrome critical region gene; DNA, deoxyribonucleic acid; DOTMA, 1,2-Di-O-octadecenyl-3-trimethylammonium propane;fied internalizing scFvs that target tumor sphere cells; HCV, hepatitis C virus; iNOP, nanotransporter interfering nanoparticle-7;ked nucleic acid; LPH, liposome-hyaluronic acid; 2′-Me, 2′-methyl; miRNA, microRNA; 2′-MOE, 2′-methoxyethyl; NLE, neutralleotides; ODN, oligodeoxynucleotides; PEI, polyethyleneimine; PNA, peptide nucleic acids; PU, polyurethane; RNA, ribonucleichain variable fragment; siRNA, small interfering RNA; TEPA-PCL, dicetyl phosphate-tetraethylenepentamine-based polycation1000 succinate; UTR, untranslated regions; VEGF, vascular endothelial growth factor.C 865), Chicago 60612-7231, USA, Tel.: +1 312 996 2253; fax: +1 312 996 2784.

ghts reserved.

963Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

1. Discovery and action of miRNAs

MicroRNAs (miRNAs) are one of a growing number of non-codingRNA molecules that act within a cell [1,2]. Noncoding RNAs includemiRNA, small interfering RNA (siRNA), ribozymes, among others [3].BothmiRNA and siRNA interact with messenger RNA (mRNA), typicallymarking the mRNA for degradation as will be described in this review.Ribozymes are RNA molecules that bind to mRNA (or other RNAs) andcatalyze the degradation of the RNA with no enzyme [4]. Ribozymesare the only RNAs in this class known to act independent of proteinsas catalysts to exert their action. Herein, we focus on microRNAs as anovel class of therapeutic molecules with an emphasis on the methodsused for delivery.

Discovered little more than a decade ago [5,6], miRNAs, small(20–24 nt) single stranded endogenous RNAs, are a class of post-transcriptional gene regulators. Since miRNA activity was confirmed inanimals [7,8] and humans [9,10], miRNAs have attracted significant at-tention due to their critical regulatory impact on biological functionsin development, cell proliferation, differentiation, apoptosis, andmetab-olism [11–15]. The primary miRNA precursor, pri-miRNA, is first tran-scribed as capped, polyadenylated RNA strands that form doublestranded stem-loop structures (Fig. 1). These pri-miRNAs are processedin the nucleus by DGCR8 and ribonuclease Drosha into 70 to 100 nt longhairpin structures, called pre-miRNAs. Pre-miRNAs are then exported tothe cytoplasm by Exportin-5 and further processed by the RNAse Dicerinto approximately 22 nt double-stranded miRNA duplexes [16–18].The miRNA duplexes enter the miRNA-induced silencing complex(miRISC) and the Argonaute protein in the miRISC unwinds the maturestrand from the passenger strand [19]. The mature strand is retained inthe miRISC while the passenger strand is released and subsequentlydegraded.

Themajority of miRNA activities [11] are proposed to be through in-teractions with the miRISC complex (Fig. 2). Guided by the mature

Fig. 1. Schematic representation of the biogenesis of miRNA. From genomic DNA, RNA istranscribed either as (1) an independentmiRNA sequence or (2) an intron that is removedfrom mRNA. Following loop formation, the pri-miRNA is processed by DROSHA into pre-miRNA. The pre-miRNA is exported from the nucleus by Exportin5 before maturation byDicer. Upon loading into the RISC complex, themiRNA duplex is unwound and thematurestrand (red strand) retained within the miRISC complex. The passenger strand (greenstrand) is degraded. (For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)

miRNA strand, miRISC binds target mRNAs, and thereby induces trans-lational repression by blocking mRNA translation and degrading themRNA [20,21]. This binding is typically to the 3′ untranslated region(UTR) of mRNA; however, 5′-UTR binding and action is also reported[20,21]. Much like siRNA and siRISC recognition ofmRNA,miRISC recog-nition of mRNA can be through perfect base pairing of the miRNA withthe mRNA strand. But, the binding of miRISC to target mRNAs does notrequire perfect pairing, which is one factor allowing one miRNA strandto recognize an array of mRNA. Although this promiscuity is thought tobe a unique feature of miRNA compared to siRNA (Table 1), there is sig-nificant overlap in the mechanism of miRISC and siRISC function [22].There is such overlap that miRNA delivery, particularly when double-stranded approximately 21 nt sequences are delivered, that the termi-nology si(mi)RNA or simply siRNA is recommended by some [23].Although the similarity in at least one mechanism of action may betrue, we propose the difference between siRNA andmiRNA to be the in-tent to knock down an individual mRNA with siRNA and an array ofmRNAs with miRNA. In that, the delivery may converge, but the broadaction desired is the primary difference between the two types ofsmall non-coding RNA.

ThemiRISC complex binding tomRNA can either repress or promotetranslation, although the latter appears to be rare [24,25]. Degradationof mRNA occurs primarily when prefect base-pairing is present anddeadenylation and destabilization of mRNA that is not perfectly base-paired appears to predominate. Even when mRNA is not degraded, ini-tiationmay be repressed directly or themRNA sequestered in regions ofthe cell (e.g., P-bodies) where low protein production takes place. If ini-tiation takes place, the ribosomes stall during translation and may dropoff the mRNA strand resulting in truncation of the protein. Also, prote-ases are recruited that degrade the protein as translation is underway.Through a combination of these mechanisms, miRNA has significant ef-fect on the proteins produced by cells (Fig. 2).

Extracellular action is directed through exosome formation [26] andnuclear actions have been proposed [27–35]. The small RNAs have beenimplicated in direct interactions with DNA, particularly promoter re-gions, or alterations of the chromatin structure thus altering finalmRNA production. Much still remains to be learned about miRNA bio-synthesis and actions, but sufficient knowledge has been gained partic-ularly in the expression ofmiRNAs in varying tissues and diseases. Aswestart to understand how altered miRNA expression can be associatedwith or even cause diseases in humans, including cancer, central ner-vous system (CNS) disorders, infectious diseases, cardiovascular dis-eases, metabolic disorders, and autoimmune diseases [36–40], newopportunities for developing novel therapeutics targeting miRNA dys-regulation will emerge. To achieve the clinical reality, an understandingof how miRNA therapeutics can act is needed.

2. MicroRNA therapeutic approaches

Using high-throughput techniques including miRNA microarrays,uniquemiRNAs expression profiles are being confirmed tomediate crit-ical pathogenic processes in human disease [38,41,42]. These findingsare rapidly being translated into recommendations targeting miRNAdysregulation. To translate these discoveries to viable therapies, appro-priate, stable molecules must be designed and chemically modifiedpoly(nucleic acids) are typically utilized for miRNA therapy. In additionto stabilization, chemical modifications can also be used to augment thespecific binding of miRNA to the miRISC complex [43,44]. The currentgeneration ofmiRNA therapeutics all utilize one ormore of the chemicalmodifications developed over the last 20 years.

2.1. Chemical modification to RNA for miRNA therapy

Naked RNA (Fig. 3) has a very short half-life in blood for several rea-sons, primarily being degraded by the abundant ribonucleases presentin blood stream [45]. Thus, thefirst generation antisense phosphodiester

Fig. 2.Mechanisms of natural miRNA-mRNA action. The roles of miRNA in protein production have been proposed to be throughmany routes. The majority of the actions inhibit proteinproduction, but several modes of promotion of protein production have also been proposed. Binding of miRNA to DNA may promote transcription directly or through the recruitment ofother factors. In addition, translation can be promoted by 5′UTR binding. Inhibitory action is likely through a combination of themechanisms shownwhich include (1) direct mRNA deg-radation, (2) deadenylation, (3) initiation repression, (4) ribosomal stalling, (5) ribosomal drop-off, (6) co-translational degradation, and (7) translation repression followingDNAbinding.

964 Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

oligodeoxynucleotides (ODNs), without further chemical modification,were proven to be ineffective in therapeutic applications [46]. Thestability of the antisense sequences is augmented using chemicallymodified oligonucleotides [47,48]. Chemical modifications (Fig. 3) in-clude phosphorothioate containing oligonucleotides [49], 2′-O-methyl-(2′-O-Me) or 2′-O-methoxyethyl-oligonucleotides (2′-O-MOE) [50],locked nucleic acid (LNA) oligonucleotides [51], peptide nucleic acids(PNA) [52], fluorine derivatives (FANA and 2′-F) and other chemicalmodifications [53]. Many of these modifications have evolved previousto the discovery of miRNA first with antisense therapies in the 1980sand 1990s [54] then with siRNA therapies [55–57].

Phosphorothioate ODNs were developed attempting to overcomethe stability issue, in which one of the non-bridging oxygen in the phos-phate group is replaced by sulfur [58]. This modification dramaticallyincreases nuclease resistance for parenteral administration, elicitingsufficient RNAse H activation for the target RNA cleavage and desiredbinding to cellular and serum proteins for uptake, absorption and distri-bution [49,59]. However, limitations include relatively short in vivo half-life, low binding affinity to RNA, and nonspecific inhibition of cellgrowth still hinders the application of phosphorothioate ODNs fortherapeutic intervention [60].

The introduction of 2′-O-methyl group to the ribose moiety in aphosphorothioate oligoribonucleotides increases binding stability andreduces the non-specific inhibitory effects on cell growth [50]. Similarly,2′-O-methoxyethyl modification markedly protects phosphorothioateoligoribonucleotides from nuclease degradation and will likely slowdown oligoribonucleotides clearance from tissues and cells. The

Table 1General properties of commonly described poly(nucleic acid) therapeutics.a

Property miRNA siRNA

Size ~20 bp ~20 bpDNA or RNA RNA RNAStructure Single or double stranded Double strandProteins directly affected Many OneProtein production change Downregulated or upregulated Downregulat

a Although these general properties are typically true, there are exceptions to the many of t

greater stability brought in by 2′-O-methoxyethyl modificationand the phosphorothioate backbone should allow for less frequentdosing and may avoid the need for continuous infusions [49].

Further utilizing modifications at the 2′-oxygen, LNAs are RNA ana-logs with the ribose moiety chemically locked by a bridge connectingthe 2′-oxygen and 4′-carbon in a RNA-mimicking N-type (C3′-endo)conformation [61]. Miravirsen or SPC3649 (Santaris Pharma, Hørsholm,Denmark), an LNA-based therapeutic used to treat hepatitis C virus(HCV) infection, progressed to Phase II clinical trials in 2010. Miravirsenefficiently suppresses HCV genotype 1a and 1b infections when admin-istered to chimpanzees, with no evidence of apparent viral resistance orside effects [62]. The promisingMiravirsen clinical resultswhere a dose-dependent decrease in HCV RNA without evidence of resistance [63]was observed suggests utilization of the LNA chemistry to developmiRNA therapeutics for the treatment of other diseases.

Peptide nucleic acids (PNA) are uncharged oligonucleotides analogsin which the sugar–phosphodiester backbone of DNA/RNA has been re-placed by an achiral structure consisting of N-(2-aminoethyl)-glycineunits. PNAs exhibit high specificity and stability without generatingunwanted toxicity [64].Most importantly, PNAs frequently donot requireassistance of delivery from transfection reagents due to the lack of charge.In addition, cell penetrating peptides can be linkedwith PNAs to enhancedelivery using standard peptide chemistry [43,65]. Naked PNA-basedanti-miR-155 specifically inhibited miR-155 in cultured B cells as wellas in mice. However, very high dosage (50 mg PNA/kg/day for 2 days inmice) was required for efficiency and the untargeted mode of deliveryposes a challenge for systemic administration [64].

Ribozyme Antisense RNA/DNA Gene therapy

≥10 bp ≥20 bp N1 kbpRNA RNA/DNA DNA

ed Single stranded Single stranded Double strandedOne One One

ed Downregulated Downregulated Upregulated

he attributes for specific cases.

Fig. 3. Chemistry and structure of miRNA therapeutic molecules. RNA molecules (RNA) has several bonds that have been modified for stability including the alpha oxygen of phosphate(red), 4′ hydrogen (green), 2′ hydroxide (orange), 4′ hydrogen (purple), and the 1′ amide linkage to the base (blue). Peptide nucleic acids (PNAs) substitute a peptide bond at the 1′ amidelinkage to the base. Phosphorothioates, methylphosphonates, and boranophosphates substitute sulfur, methyl, and a borano group, respectively, for the alpha-oxygen of the phosphate.Lockednucleic acids (LNAs) add a secondary linkage between the 4′ carbonand the2′hydroxidewhile 2′-O-(2-methoxyethyl)- (2′-O-MOE), 2′-O-methyl- (2′-O-Me), and 2′-fluoro- (2′-F)RNAs substitutes less reactive groups for the 2′ hydroxyl group. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

965Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

Depending on the expression status of the targetmiRNA,miRNA ther-apeutic approaches can be separated into two categories: (1) miRNA in-hibition therapy when the miRNA is overexpressed and (2) miRNAreplacement therapy when the miRNA is repressed. These methods canbe accomplished with small RNAs directly delivered (Fig. 4) or by moreconventional gene therapy (Table 1) approaches where plasmids orvirus are delivered to express the therapeutic molecules. The gene thera-py approach will not be further discussed as the limits of delivery areequivalent to those of other gene therapy approaches and the activity issimilar to the systems described herein.

2.2. miRNA inhibition therapy

When an upregulated miRNA contributes to disease pathology,miRNA inhibition therapy can be used to block the miRNAs repressionof protein expression (Fig. 4). Severalmethods have been used to inhibitmiRNA repression of protein expression, but each involves the disrup-tion of the miRISC complex. The most straightforward method utilizesoligonucleotides complementary to the miRNA mature strand, anti-

miRNA oligonucleotides (AMOs), to inhibit interactions betweenmiRISC proteins and the miRNA or miRISC and its target mRNAs. Thisapproach mimic the idea of anti-sense DNA and RNA toward themRNA strand preventing translation of the mRNA [66,67]; however,the AMO is designed to bind the miRNA (not the mRNA), prevent deg-radation of the mRNA, and allow the mRNA to be translated. In orderto achieve effective inhibition, chemical modifications are often usedto enhance selective hybridization with the endogenous miRNAs.

Similar to AMO therapy, miRNA sponges seek to occupy the maturemiRNA, but instead of delivering short sequences complementary to themiRNA mature strand, DNA sequences, typically plasmids, expressingmiRNA target mRNA in high copy number are transiently transfectedin cells. This approach has recently been reported as similar to naturalphenomena occurring in mammalian cells where circular RNA actsas a miRNA sponge inhibiting activity [68]. Those transcripts containmultiple binding sites to the miRNA of interest and saturate the miRISCcomplex repressing the activity toward natural mRNA [23]. Using thisapproach, miR-9 sponge, previously identified as having metastasis-promoting function in breast cancer cells, inhibited breast cancer

Fig. 4. Therapeutic strategies for miRNA activity. Delivered DNA or RNA (green) can act to augment, mimic miRNA activity, or prevent activity of miRNA. MicroRNA can be produced byplasmid DNA (miRNA gene therapy) that is introduced into the cell, or by introducing agents to increase by translation promoters or cofactors, or delivered directly (in the forms of pri-,pre-, ormaturewith orwithout passenger), i.e.miRNAmimic. To prevent the activity of expressedmiRNAs,masks that competewith the binding site onmRNA inhibitmiRNAbindingwithmRNA. In addition, binding of miRNA with the RISC complex may be inhibited with a sponge or a single miRNA binding anti-miRNA. Finally, translation may be directly repressed at theDNA level. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

966 Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

metastasis formation [69]. However, due to limited control and homo-geneity of transcripts expression, miRNA sponges could lead to seriousunwanted side effect and thus is better suited for studying miRNAsfunction in cell culture assay in vitro.

In contrast to occupying the miRISC complex with decoys, miRNAmasks were designed to complement the miRISC binding sites in the3′ UTR of the target mRNA [70]. In zebrafish, this gene-specific,miRNA-interfering strategy prevented the miR-430 repression oftransforming growth factor-β signaling [71]. The most significant ad-vantage of this strategy lies in that signaling pathways regulated by amiRNA can be selectively blocked by choice of sequence, but this strat-egy also requires the clear profile of miRNA target genes. Further, themiRNA masks are designed to have specificity for specific mRNA, andthus lose the ability to return protein production for many proteinsother than the specific protein that's mRNA is masked.

2.3. MicroRNA replacement therapy

miRNA replacement therapy supplements the lowered level ofmiRNAs with oligonucleotide mimics containing the same sequence asthe mature endogenous miRNA, known as miRNA mimics. In order toachieve the same biological function as the naturally produced miRNA,mimics should possess the ability to enter the RISC complex and affectmiRNA target mRNAs. Theoretically speaking, a single strand RNAmol-ecule containing the same sequence, as the mature miRNA would workas amiRNAmimic. However, double strandedmiRNAmimics composedof a guide strand and a passenger strand have 100 to 1000 fold higherpotency compared with single stranded miRNA mimics [72]. Theguide strand contains a sequence identical to the mature miRNA andthe passenger strand sequence is complementary to the maturemiRNA. For example, intratumoral injection of cholesterol-conjugatedmiR-99a mimics significantly inhibited tumor growth in hepatocellularcarcinoma-bearing nude mice [73]. In addition to the miRNA mimicshaving identical sequence as the endogenous mature miRNA, syntheticmiRNA precursor mimics with longer sequence ranging from just a fewadditional nucleotides to full length pri-miRNA have been proposed[74]. Since pri-miRNA is processed in the nucleus, significantly different

strategies would be mandated for nuclear targeting of pri-miRNA thatwould not be necessary for pre-miRNA or mature miRNA. In any case,choosing the appropriate materials to augment the delivery of themiRNA appears to be one of the most important factors influencingthe activity of the miRNA.

3. Synthetic materials for miRNA and anti-miRNAoligonucleotide delivery

Syntheticmaterials have demonstratedpotential as effective carriersfor DNA and siRNA [75–79]. Undermost circumstances, synthetic mate-rials are cationic and condense negatively charged poly(nucleic acid)sthrough electrostatic interactions. Synthetic delivery systems have con-siderable advantages over viral-based vector due to the control of theirmolecular composition, simplified manufacturing, modification andanalysis, tolerance for cargo sizes, and relatively lower immunogenicity[80]. However, synthetic systems have relatively lower efficiency com-pared to viral vectors. Relative efficiency has been greatly improvedby modifying particle size and surface properties to achieve specificbiodistribution in vivo. Further, conjugating small molecule ligands topolymeric carrier targeting diseased cell surface specific receptors hasbeen widely used to augment cell uptake in appropriate cells.

The administration of miRNA expressing vectors, i.e. DNA sequencesthat would be transcribed into pre-miRNA, falls within the purview ofDNA delivery, which requires not only nuclear localization but also fur-ther processing by RNAmachinery. In addition to the general barriers tomiRNA delivery, delivery of miRNA expressing DNA resulting in highlevels of expression which can saturate exportin-5 resulting in off-target effects [81]. Although this too is miRNA delivery (some wouldconsider this the only type of miRNA delivery), we focus our discussionto research that used synthetic miRNA molecules, such as pre-miRmimics, mature miRNA mimics, and AMO [82,83].

3.1. Lipid-based delivery system

Liposomes are one of themost commonly used transfection reagentsin vitro; however, safe and efficacious delivery in vivo is rarely achieved

967Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

due to toxicity, nonspecific uptake, and unwanted immune response[84]. Much of the nonspecific response and toxicity is directly linkedto the positive charge on the surface of the particles necessary for thebinding of oligonucleotides. Currently, there are several commerciallyavailable lipid-based delivery systems (Table 2) andmost investigationsvalidating miRNA targets in vitro have used such systems. In recentyears, significant effort has been dedicated to modifying the composi-tion and chemical structure of liposomes, and some success has beenachieved in delivering siRNAs [85].

Pre-miR-133b containing DOTMA:cholesterol:TPGS lipoplexeswere prepared by adding pre-miR-133b to the empty liposomes.The in vitro transfection efficiency and in vivo biodistribution oflipoplex formulations were compared with siPORT NeoFX transfec-tion agent. In vitro, the lipoplexes transfected pre-miR-133bmore ef-ficiently than siPORT NeoFX, a commercially available lipid-basedagent, in A549 non-small cell lung cancer cells. Approximately 30%of Cy5 labeled oligodeoxynucleotides lipoplexes accumulated inthe lung, which was 50-fold higher than the NeoFX formulation.The mature miR-133b level in lungs following i.v. administration ofpre-miR-133b containing lipoplexes was approximately 52-foldhigher than untreated mice [88]. However, no further therapeuticeffect on inhibiting tumor growth was reported. Using DDAB:choles-terol:TPGS based lipoplexes, head and neck squamous cell carcinomawas treatedwith systemic administration of pre-miR-107 lipoplexes.

Table 2Representative recent nonviral gene delivery systems for miRNA therapeutics.

Material Disease model

Naked oligonucleotidesN/A Metabolic diseaseAptamer-ologonucleotide Ovarian cancer

Lipid-based delivery system: commercially availablesiPORT Non-small cell lung cancer cells;

LymphomaMaxSuppressor Non-small cell lung cancerLipoTrust Colorectal tumors

Lipid-based delivery system: newly explored98 N12−5 Fat and cholesterol metabolismDOTMA:cholesterol: TPGS lipoplexes Non-small cell lung cancer cellsDDAB:cholesterol:TPGS lipoplexes Head and neck squamous cell carcinomaGC4 scFv-targeted LPH nanoparticles Murine B16F10 melanomacRGD-targeted LPH nanoparticles AntiangiogenesisiNOP-7 Fat and cholesterol metabolismDC-6-140-DOPE-Choleterol liposomes Lung cancerSolid lipid nanoparticles Lung cancer

Polyethyleneimine (PEI)PEI Colon carcinomaPolyurethane-short branch PEI Lung adenocarcinoma;

GlioblastomaRVG-coupled PEI Neuron diseases

DendrimersPoly(amidoamine) Glioblastoma

Poly(lactide-co-glycolide) (PLGA)Nonaarginine-modified PLGA nanoparticles KB cellsPenetratin-modified PLGA nanoparticles Pre-B cell lymphoma

Natural polymersAtelocollagen Colorectal cancer;

Lymphoma;Prostate

ExosomesExosome Breast cancer

Inorganic materialsDisialoganglioside-targeting silica nanoparticles NeuroblastomaRGD modified ultrasmall magnetic nanoparticles Breast cancer

Tumor growth was reduced by 45.2% compared to lipoplexes con-taining pre-miR-control [93]. Although effective, the inclusion ofcharged lipids on the surface of the liposomes suggests that in-creased clearance of the liposomes would be observed. Additionally,limited information is presented on the stability of the interactionfollowing injection into biologic fluids where the high salt concen-tration and proteins may greatly alter the ionic complexes.

Polycationic liposome-hyaluronic acid (LPH)nanoparticles have alsobeen described by several investigators [112]. As applied to miRNA de-livery, a tumor targeting GC4 single-chain antibody fragment modifiedLPH (scFv-LPH) nanoparticles systemically co-delivered siRNA andmiR-34a into experimental lung metastasis of murine B16F10 melano-ma. The scFv-LPH nanoparticles encapsulating combined siRNAs againstc-Myc, MDM2, and VEGF and miR-34a decreased the metastasis tumorgrowth to about 20% of the untreated control. When treated withscFv-LPH nanoparticles only containing combined siRNAs or miR-34a,the reductionwas about 30 and 50%of the untreated control, suggestingthe effects are through different mechanisms. The advantage of such asystem lies in the potential to deliver siRNA and/or miRNA togetherto simultaneously target several different oncogenic pathways [94].Using the same liposome-polycation-hyaluronic acid nanoparticle con-jugatedwith cyclic RGD, an integrin-binding tripeptide (cRGD-LPH) de-livered anti-miR-296 efficiently interrupting blood vessel formulationand endothelial cell migration in human umbilical vein endothelial

Targeted miRNA Therapeutic approaches Route Refs

miR-122 Inhibition Systemic [61,62,86]let-7i Replacement Systemic [87]

miR-133b;miR-34a Replacement Systemic [88,89]

miR-34a/Let-7 Replacement Systemic [90]miR-143/miR-145 Replacement Intratumoral/Systemic [91]

miR-122 Inhibition Systemic [92]miR-133b Replacement Systemic [88]miR-107 Replacement Systemic [93]miR-34a Replacement Systemic [94]miR-296 Inhibition Subcutaneous [95]miR-122 Inhibition Systemic [96]miR-7 Replacement Systemic [97]miR-34a Replacement Systemic [98]

miR-145/miR-33a Replacement Intratumoral/systemic [99]miR145 Replacement Systemic; Intracranial [100]

miR-124a Replacement Systemic [101]

miR-21 Inhibition Cell culture [102,103]

miR-155 Replacement Cell culture [104]miR-155 Replacement Systemic [105]

miR-34a;miR-135b;miR-16

Replacement;Inhibition;Replacement

Intratumoral;Intratumoral;Systemic

[106–108]

let-7a Replacement Systemic [109]

miR-34a Replacement Systemic [110]miR-10b Inhibition Systemic [111]

968 Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

cells by specific uptake throughαvβ3 integrin. Anti-miR-296 containingcRGD-LPH nanoparticles exhibited themost significant inhibition of cellproliferation and microvessel formation in matrigel plugs subcutane-ously implanted in mice [95]. Even though these targeted miRNA ther-apeutic delivery systems showed enhanced activity in tissue cultureandmousemodels, they do not guarantee the same result in human be-ings. Like most targeted delivery systems, the accumulation of the par-ticles is limited by the flow of particles into the diseased tissue, whichtypically is limited to less than 5% of the total particles administered.The targeting moieties only allow particles to be taken up by cells theyreach and cannot generally increase the amount of particles reachingthe diseased cells.

To overcome disruption of the ionic interactions thatstabilize most liposome-miRNA interactions, dicetyl phosphate-tetraethylenepentamine-based polycation liposomes (TEPA-PCL) werecombined with cholesterol-grafted miR-92a, an angiogenesis regulat-ing miRNA, for cancer anti-angiogenesis therapy. The miR-92a/TEPA-PCL complex showed significant inhibition on endothelial cell tubeformation at 200 nM in vitro [113]. Although complicating the pro-duction and design of the therapeutic molecules, combinations of de-livery strategies hold the greatest promise to overcoming the currentlimitations of the delivery platforms that are available.

Even though cationic lipids are the most widely used for lipid-basedsystem, their in vivo application is frequently limited by toxicity [84].Aiming to overcome this hurdle, a neutral lipid emulsion (NLE),MaxSuppressor in vivo RNALancerII (BIOO Scientific, Inc.) was devel-oped. MiR-34a and Let-7 were delivered with a NLE to treat non-smallcell lung cancer (NSCLC). Unlike cationic lipids, NLE preferentially accu-mulated in the lungs rather than the liver. The treatment with miR-34aor Let-7 significantly decreased lung tumor burden to approximately40% of the mice treated with miRNA controls and the expression levelof miR-34a and let-7 in lungs was also significantly higher than groupstreated with miRNA mimic controls [90]. These findings demonstratethe potential of developing miRNAs therapy formulations with NLE asnovel therapies for lung cancer patients. These particles, however, arelimited in the diseases for which they could be applied as they rely onthe passive accumulation in the lung. Strategies to overcome this limita-tion should be sought, but few methods have overcome the passive ac-cumulation of particles in the liver, lungs, or spleen.

In addition to beingused directly as delivery carriers, lipidswere alsoused as modification agents using a dendrimer-like structure of lysinewith lipid surfaces, referred to as iNOP-7. The delivery of apoB siRNAwith iNOP-7 efficiently silenced apoB mRNA, decreased apoB proteinlevels in liver and plasma, and lowered total plasma cholesterol [114].Similarly, anti-miR-122, deliveredwith iNOP-7 (2 mg/kg) administeredon three consecutive days, effectively silenced 83.2 ± 3.2% of miR-122expression, which was accompanied by regulating miR-122 targetgenemRNA expression in liver and lowering of plasma cholesterol [96].

There is no doubt lipid-based materials will be one of the mostimportant candidates for further exploitation of miRNA therapeuticsformulation. As one of themost widely studiedmaterial for pharmaceu-tical delivery, the knowledge accumulated forms a solid foundation forany further exploration. However, lipid-based systems are not theonly option and other alternativematerials might be needed for specificdisease and offer unique advantages.

3.2. Polyethylenimine (PEI)-based delivery system

Polyethylenimine is one of the most widely used and studied poly-mers for gene delivery. In the physiological milieu, PEI is positivelycharged due to protonation of the amine groups and thus can be usedto condense nucleic acids. Cationic polyplexes formedbyPEI and nucleicacids typically retain a net positive charge (ζ-potential) promotinginteractionswithnegatively chargedpolysaccharides on the cell surface.Once interacting with the cell surface, it is proposed that the complexesundergo endocytosis. The polyplexes escape from the endosome

through so called ‘proton sponge’ effect where PEI causes influx of pro-tons andwater and endosome swelling and eventually disruption to re-lease the polyplexes to cytoplasm [115]. Using PEI, both DNA and siRNAhave been successfully delivered in animal models in vivo, as has beencomprehensively reviewed elsewhere [80,99,116]. Many of the samelimitations that have limited the translation of PEI to the clinic are alsopresent for miRNA delivery although PEI remains in the pipelinefor translation [117,118] due to the augmented activity compared tomany other polymeric systems.

Using a polyurethane–short branch polyethylenimine (PU–PEI) as acarrier, miR-145 was delivered to treat cancer stem cell (CSC) derivedlung adenocarcinoma (LAC). The LAC–CSC xenograft tumors did not re-spond to the combination of ionizing radiation (IR) and cisplatin duringthe 30-day experimental course; however, PU–PEI-bound miR-145delivery moderately reduced tumor growth. Most importantly, themiR145 delivery combined with IR and cisplatin led to significanttumor growth inhibition [100]. When administered to orthotopic CSC-derived glioblastoma tumors, intra-cranially delivered PU–PEI-miR-145 significantly suppressed tumorigenesis. When used in combinationwith radiotherapy and temozolomide, synergistic effects and improvedsurvival rate were achieved [119]. The significant inhibitory effectof PU–PEI-miR-145 on lung adenocarcinoma and glioblastoma CSC-induced tumors demonstrated the potential of miRNA therapy in over-coming tumor chemoradioresistance, preventing cancer relapse andachieving cancer eradication.

Beyond traditional delivery approaches, PEI-based systems havebeen modified for transport across the blood–brain barrier (BBB). TheBBB is the most significant physiologic obstructions of systemic drugor gene delivery to the brain parenchyma and central nervous system(CNS) [120]. Using a short peptide derived from rabies virus glycopro-tein (RVG), the PEI–RVG bound specifically to nicotinic acetylcholine re-ceptors on neuronal cells. RVG was coupled to PEI via disulfide bond(RVG–SSPEI) to deliver miR-124a, a neuron-specific miRNA that canpotentially promote neurogenesis [101,121]. To overcome the size lim-itation of PEI vector transport across the BBB, mannitol was used topermeablized the BBB. After administered, much higher accumulationof miR-124a in the brain was observed in the RVG-mediated SSPEI de-livery group compared to that in the miR-124a/SSPEI group by trackingthe Cy5.5 labeled miR-124a [101]. However, the functional activities ofmiR-124a to promoting neurogenesis were not tested. Themodificationof PEI usingRVGdecreased the toxicity associatedwith PEI and achievedremarkable targeted delivery to neuronal cells. The RVG–SSPEI could bea useful system to delivermiRNA therapeutics for the treatment of braindiseases. Although this system did show greater accumulation in thebrain, the limitation of using permeabilizing agents limits the utility.Combination of delivery strategies that each improves the activity ofthe miRNA has great potential, but the complexity of the systems attimes can counterbalance the improvements [122].

3.3. Dendrimers

Dendrimers have attracted significant attention in gene delivery dueto their defined architecture and a high ratio of surface moieties to mo-lecular volume [123]. Dendrimers have been successfully used to deliverDNA and RNA; however, few applications to miRNA delivery have beenreported. Co-delivery of anti-miR-21 and 5-fluorouracil (5-FU) to U251glioblastoma cells using poly(amidoamine) (PAMAM) dendrimer in-creased apoptosis of U251 cells markedly. Migration of tumor cellswas decreased compared with cells that were only treated with 5-FU[103]. Anti-miR-21 delivered by PAMAM increased the chemosensitivityof humanglioblastoma cells to taxol [102]. Although not in vivo, this sug-gests that dendrimers may be amenable for in vivo miRNA delivery.Dendrimers are capable of binding miRNA and aiding in the entry intocells, but the entry is non-specific in nature. The alternate mechanismof cell entry that should not be overlooked as this has the potential to

969Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

protect miRNA by avoiding the endosomal and lysosomal compart-ments [124].

3.4. Poly(lactide-co-glycolide) (PLGA) particles

The majority of the previously described polymers focus on theability to condense miRNA and non-specifically enter the cell whilepoly(lactide-co-glycolide) (PLGA) focus on these factors but also allowthe release of miRNA over time. PLGAs are a family of water-insolublepolymers that have been widely used in biomedical applications overthe last half-century [125]. PLGA polymers can be readily formed intomicroparticles and nanoparticles, which entrap biologically active mol-ecules [126,127]. PLGA particles overcome several limitations facingcurrent miRNA therapy because they can protect nucleic acids fromdegradation, achieve high loading capacities, and affordmultiple surfacemodifications so as to generate potentially favorable pharmacodynam-ics [128]. Successful delivery of DNA and RNA has been achieved andsome ongoing clinical trials showed great potential [75].

Using a miR-155 Cre-loxP tetracycline-controlled knockin mousemodel, pre-B-cell tumors were dependent on high miR-155 expressionwhere withdrawal of miR-155 using doxycycline caused rapid tumorregression. Systemic delivery of anti-miR-155 peptide nucleic acids(PNAs) using PLGA polymer nanoparticles exhibited enhanced deliveryefficiency and achieved therapeutic effect. The surface of the nano-particles was modified with penetratin, a cell-penetrating peptide[105]. Although their previous studies have shown that PLGA NPs mod-ified with nona-arginine can effectively deliver anti-miR-155 PNAs toKB cells and achieve therapeutic miRNA inhibition [104], pre-B cellspreferentially took up nanoparticles coated with penetratin (ANTP-NP). The ANTP NPs preferentially accumulated in tumor tissue com-pared with other organs and tissues through enhanced permeabilityand retention effect (EPR) [129]. The pre-B cell tumors had an approxi-mately 50% decrease in growth relative to control-treated tumors aftersystemic delivery of 1.5 mg/kg anti-miR-155 PNAs loaded in ANTP–NPfor 5 days, which was approximately 25-fold less than naked anti-miRdosage needed [105]. There was need, in this case, to protect the PLGAparticle using steric stabilization, i.e. PEGylation, and also add a cell pen-etration enhancer. PLGA particles are typically non-specifically clearedand the PEGylation diminished the ability of the particles to entercells. This type of particle is readily adaptable, but still does not have sig-nificantly more than 5% accumulation in the diseased organ due to pas-sive accumulation.

One of the greatest advantages of PLGA polymers is their adaptabil-ity and ability to load multiple cargos. MRI-detectable nanoparticleswere synthesized using PLGA and loaded with perfluoro-1,5-crownether (PFCE) for 19 F-MRI tracking. These PLGA particles also containedprotamine sulfate for miRNA condensation. The nanoparticles accumu-lated in the endosomal and lysosomal compartments and efficiently re-leased miR-132, inducing potent pro-survival activity in endothelialcells (ECs) transplanted in vivo. There was also a strong enhancementblood perfusion in ischemic limbs relatively to control. Important tothe future development of miRNA is the assertion by the authors thatRNA degradation occurs in or around the endosomal and lysosomalcompartments [130,131]. If this proves true, much of the design criteriafor miRNA will alter. The fact that Ago2 resides in the membrane of theendolysosomal compartment could explain the better performance ofNP170-PFCE formulation [132]. The future development of miRNA(and possibly siRNA) will rely on the rational design of systems thattake advantage of the complex biology of the disease and moleculebeing delivered.

3.5. Other nonviral miRNA therapeutics delivery system

In addition to synthetic polymeric material as delivery system, natu-rally occurring polymers, such as chitosan, protamine, atelocollagen,and peptides derived from protein translocation domains have also

attracted significant attention [133,134]. However, their application isusually limited by immunogenicity [135]. Atelocollagen exhibits limitedimmunogenic response because it is purified from pepsin-treated type Icollagen, during which antigenicity conferring telopeptide is removed[136]. Similar to other materials used for gene delivery, atelocollagenis positively charged and can condensate oligonucleotides by formingnanoparticles through electrostatic interactions. Due to its successfulapplication in delivering DNA and siRNA [137], several research groupsattempted to use it for miRNA therapeutics administration. MiR-34a/atelocollagen complexes administered intratumorally into the subcuta-neous spaces around human colon cancer mice xenograft suppressedtumor cell growth in vivo [106]. Using the same approach, an LNAmiR-135b inhibitor was administered subcutaneously to human lym-phoma mice xenograft and tumor growth was significantly inhibitedcomparing with the negative control groups at 13 days post tumorinoculation [107]. In addition, miR-16/atelocollagen complexes weredelivered systemically to a prostate cancer mouse xenograft model.Atelocollagen not only delivered miR-16 to bone tissue but also in-hibited metastatic tumor growth in bone tissues without causing no-ticeable side effects [108]. The choice of natural materials may havegreat advantages–specifically in terms of biologic acceptance and abilityto mimic natural processes–if the properties allow sufficient efficiency.

Using another biologically derived molecule, protamine, miR-29bwas delivered to human mesenchymal stem cell (hMSC). Complexesformed were around 34 nm with a zeta potential +27 mV. ThemiR-29b/protamine complexes were readily taken up by hMSC posttransfection while lipoplexes exhibited very little transfection. In addi-tion, the inhibition of miR-29b/protamine complexes on cell prolifera-tion was almost negligible, which is a critical factor for transfectingstem cells since the cell number is a critical factor for their use as ther-apeutics. MiR-29b/protamine complexes were not only able to inducethe expression of osteoblast differentiation markers, their expressionlevel also showed a dose response to increasing concentration of miR-29b/protamine formulation [138]. Using biologically derived proteinsmay allow the direct use of human molecules to deliver miRNA partic-ularly when human proteins can be utilized and have significantlydecreased toxicity.

Inorganic materials have also been exploited to produce nano-particles. The inorganic materials usually have monodispersed particlesize and good optical properties. As a model miRNA, miR-130b wascoupled to gold nanoparticles through oligo(ethylene glycol) thiol(miRNA-AuNPs). The miRNA-AuNPs own good surface stability sincethemiRNAdid not showexchange activitywith a highly charged surfac-tant when assessed by gel electrophoresis. The Cy5-labeled miRNA-AuNPs was found to localize inside the cells with confocal microscopyand the cell morphology was not significantly changed after transfec-tion.MiRNA-AuNPsmediated efficient knockdown in functional lucifer-ase assay compared with the controls [139]. Another gold nanoparticlesbasedmiRNAdelivery system first functionalized AuNPswith cystamineand modified the miRNA-AuNPs with thiol- polyethylene glycol postmiRNA loading, which demonstrated high payload, low cytotoxicityand robust miRNA transfection capacity [140]. These systems, as withmany of the non-viral systems, have not been assessed in vivo. There ispotential for delivery using these systems, but there are also questionsthat need to be answered concerning toxicity and clearance.

Similarly, silica-based nanoparticles are stable and inert with biode-gradable and non-toxic properties [141]. Silica particles release thecargo through hydrolysis of the silica network. Taking advantage ofneuroblastoma's property of expressing high level of cell surfaceantigen disialoganglioside (GD2), GD2-targeting silica nanoparticleswere designed to deliver miR-34a into neuroblastoma tumors. Anti-GD2 nanoparticles containing miR-34a resulted in a significant up-regulation of miR-34a in the neuroblastoma cell lines, but not inHEK293. At day 14, 17 and 20 post-tumor establishment, anti-GD2-miR-34a-NP or anti-GD2-miR-scramble-NP were administered to NB1691luc

or SK-N-ASluc mice. Tumor sizes in NB1691luc and SK-N-ASluc mice were

Table 3Barriers and solutions to miRNA delivery (Adapted from Scholz and Wagner [154]).

Barrier/step to administration Solutions

Administration Local administrationI.v. administration

Degradation and elimination Local administrationCondensation to carrierControl of• particle size• particle shape• particle charge• steric stabilizationChemical modification

Disease accumulation Local administrationTargetingParticle size

Cellular entry Targeting ligandsNon-specific interactionsCell penetrating moieties

Endosomal/lysosomal Endosomal escape• lytic lipids• fusogenic peptides• fusogenic polymers• osmotic lysisEndosomal arrest/lysosomal inhibitionMVB targeting

Intracellular localization MVB/P-body targeting for RISC activity

970 Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

monitored by bioluminescent imaging on days 18, 21 and 25 and a signif-icant reduction in the bioluminescent intensity of the tumors was ob-served in mice treated with anti-GD2-miR-34a-NPs relative to negativecontrol. Investigation of the molecular mechanism of the tumor growthinhibition effect revealed that delivery of miR-34a using anti-GD2-NPscaused statistically significant increase in apoptosis and mediated nega-tive impact on vascularization through downregulation of MYCN proteinand upregulation of tissue inhibitor metallopeptidase 2 precursor(TIMP2) protein [110].

To more specifically target the disease sites, mediate effectivetherapeutic effect, and contain traceable label for monitoring drug dis-tribution, magnetic particles (MP) have also been proposed [142].Ultra-small magnetic nanoparticles surface modified with RGD andLNA oligonucleotides were designed as an imaging-capable particlethat can alleviate breast cancer metastasis through downregulatingmiR-10b level. BothMRI and infrared fluorescence optical imaging indi-cated that the particles preferentially accumulated in the primary tumorand lymph nodes comparingwithmuscles. In orthotopicMDA-MB-231-luc-D3H2LN tumors, particles prevented breast cancer metastasis tolymph nodes when the treatment was initiated before detectable me-tastasis formation. Surprisingly, even though the particles did not inhib-it primary tumor growth, an inhibitory effect on the growth of lymphnode metastases when the treatment was given post lymph node me-tastases formation was observed. It was proposed that miR-10b playsa context-dependent role in breast tumorigenesis, namely miR-10bcontrolled metastasis but not tumor growth in the primary tumorwhile in the lymph nodes metastases the miRNA extended to regulatetumor proliferation [111]. This finding indicates the possibility of effec-tively treating disseminated cancer by not only specifically targetingto the tumor metastases but also monitor the process by noninvasiveimaging approach. The elucidation of potential context-dependentrole of miRNAs in oncogenesis opens new door for developing anti-metastatic therapies.

4. Future outlook: challenges and opportunities

Over the last several years, miRNA research has made amazinglyrapid progress. But despite this, there is a belief that “RNAi is dead”[143] partly due to the departure from or reduction in emphasis onsiRNA and miRNA by the pharmaceutical industry [144]. Few siRNAtherapeutics and only two candidate miRNAs, SPC3649 (SantarisPharma, Horsholm, Denmark) [145–151] a miR-122 antisense lockednucleic acid, and MRX34 (Mirna Therapeutics, Inc.) [152] a liposomalmiR-34 mimic, have reached clinical trials. There is significant risk forinvestment by the pharmaceutical companies that is due to the biolog-ically challenging therapeutic molecules, the expense of making andscaling up the therapeutics, and the unproven delivery systems withclinical approval challenges [153]. The field is still in the infancy andwill take some time to mature. The costs and challenges will only be-comemore reasonable. Of all of the scientific (Table 3) and commercialchallenges, the greatest challenge to translating siRNA andmiRNA is de-livery [143].

It is now widely recognized that designing miRNA specific for a dis-ease target is not the bottleneck for RNA therapeutics, but the design ofappropriate delivery systems that will reach pharmacologic targetsin vivo is the key to miRNA development. The development of RNAitherapeutics has to be considered in terms of which cell/tissue typesavailable delivery technologies can address. A sing system will not bethe answer for all diseases. The examples are liver-directed gene knock-down [145,146,152], bladder cancer-directedmiRNA therapy [117,118],and endothelial cell-directed knockdown using Silence Therapeutics'cationic AtuPLEX lipoplex technology [155] which each reach organsthat are known to take up particles or where particles can be directlyin contact with the diseased tissue. Much progress has been gained inthe area of disease targeted drug delivery, but evenwith these advances

there are significant limits to the amount of drug that accumulates inthe diseased organs [156].

Many of the parameters necessary to control the delivery of any par-ticulate system are well established. To improve the residence time inthe circulation, a material should be greater than the renal filtrationlimit [157–159], but it should be excreted when the payload is released[160]. While in the bloodstream, particles must avoid the uptake byphagocytic cells [161,162]. The charge on a particle, typically posi-tive for RNA condensation, has dramatic influence on the cellularuptake. Most transfection reagents are highly cationic and takenup through non-specific means, but this also leads to rapidclearance. Competing with these issues is the need to cross the en-dothelial layer to reach the diseased tissue [163–167], when not ad-ministered for local response. Even when the endothelial barrier istraversed, the local extracellular matrix yields a substantial resis-tance to penetration into the tissue [163–167]. The systems de-scribed within this review address each of these issues in differentways.

The inherent properties of miRNA are a significant challenge todelivery. Molecular stability before reaching the patient is a con-cern and this is often ignored. This is coupled with the cost andcomplexity of the production of the molecules and delivery sys-tems. Once the miRNAs are administered, unprotected miRNAs arerapidly degraded by serum nucleases in bodily fluids. While chem-ical modifications protect miRNA, these molecules can impair spec-ificity and potentially introducing off-target effects [57,168]. This iscompounded by the clearance of particulate systems even whenthey adequately protect the miRNA from degradation. When parti-cles reach cells, the miRNA must separate for action to be possible.The limited separation of miRNA from the carrier has forced in-creased dose of miRNA to be active. Many current systems are fo-cusing on using targeted drug delivery strategies to increase theamount of the drug that reaches the site, but targeting has a limitedeffect at increasing the total amount of drug that reaches the targetorgan [169].

The chemical and physical properties are not the only challengesas biologic properties of miRNA present difficulties that must becomprehended to properly deliver the molecules. The property ofmiRNA being able to regulate a network of genes without perfectpairing also imposes the challenge of consistency and predictability of

971Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

miRNA therapeutics. Moreover, the presence of exogenous artificialmiRNAs brings the risk of saturation of RNA processing machinery andcause untoward side effect [81,170]. Most importantly, the expressionlevel and function of miRNAs vary not only between normal and dis-eased tissues and organs but also between disease stages [171,172].For example, a total of 28 epigenetically regulated miRNAs were identi-fied as being associated with poor patient survival when underexpressed in neuroblastoma, amongwhichmiR-340 induced either dif-ferentiation or apoptosis in a cell context dependentmanner [173]. Thisfeature is challenging because it requiresmiRNA therapeutics to achieveboth temporal and spatial control over biodistribution in vivo.

Evenwhen active, small RNAs can have nonspecific effects thatmustbe examined. Interferon response can occur following activation inter-feron stimulated genes [174–177]. When this happens, it leads to non-specific inhibition of protein synthesis coupledwith non-specific degra-dation of endogenous mRNA. Therefore, in addition to non-targetingRNA controls, one must rule out the possible interferon responsewhen validating the small RNA. This can be achieved by determiningthe expression levels of several key genes involved in the interferon re-sponse [178–181]. These responses can be controlled possibly by appro-priate delivery and design of the therapeutics.

To begin to validate the technologies, diseases with limited alterna-tives for treatment have been chosen. This is exemplified by the factthat only two oligonucleotide-based therapeutics, Vitravene® andMacugen®, have reached the market and both are for intravitreal ad-ministration, chosen due to the inherent difficulty to deliver and direneed for therapy. These products, although successful clinically, arenot considered to be successes commercially. Part of the commercialchallenge has come from the entry of competing products that are lessexpensive and similar if not more effective. The commercial success isalso limited by the expense of the products themselves. There is muchroom to improve the commercial success of the products as we learnfrom the current generation of miRNA therapeutics and delivery sys-tems. As the cost of production and scale-up of the products decreases,there will be success in this field, both commercial and clinical. Duringthe design of newmiRNA delivery system, the feasibility and cost asso-ciated with production for clinical translation should be considered.

As we develop delivery system, we must focus on choosing diseasesthat have high need for new, and potentially expensive, therapies.Unique miRNA expression profiles have been identified in untreatableor poorly treated diseases, such as cardiovascular disease, neurologicaldisorders, and autoimmune disease [37,39,182]. An array of miRNAswere found to function as gene regulators in Alzheimer's diseaseby targeting the mRNAs of amyloid precursor protein (APP), β- and γ-secretase, of which the latter two are responsible for the cleavageof APP and the generation of amyloid-beta peptide, a significanttarget for treating Alzheimer's disease [183]. Cordes and colleaguesdemonstrated that miR-145 was necessary and sufficient to inducemultipotent stem cells differentiation into vascular smooth musclecells [184]. This important finding raises the possibility of suppressingsmooth muscle hyperplasia observed in vascular injury and atheroscle-rosis using miR-145 replacement therapy. The important roles playedbymiRNAs in diseases for which conventional medicine offers no effec-tive solution indicates that focusing miRNA therapeutics developmenton those diseases is worth the effort and warrants further study.

The complexity and heterogeneity of signaling pathways involved inmany diseases, including tumor initiation, progression, metastasis andchemo/radio resistance accounts for the difficulty of developing effec-tive medicine [185]. The discovery of miRNAs might be the criticalpiece of the cancer research puzzle since one miRNAs can regulatea broad network of genes involved in different signaling pathways.This fact is both a significant advantage for miRNA compared tosiRNA and a biologic challenge as untoward effects may be aug-mented by the broad activity. Proper design of the AMO andmiRNA mimic sequences [186,187] is the key to the ability to selec-tively alter the necessary networks. This fact is also one of the

greatest challenges due to the fact that the prediction of side effectsmay be difficult.

Many miRNAs have been identified as key oncogenic targets in dif-ferent kinds of cancer, including cancer stem cells [70,188–193]. Thosefindings suggest that therapeutic intervention of supplementing atumor suppressor miRNA or inhibiting an oncomiR may indeed inducemultiple antitumor effects by simultaneously interfering with severaloncogenic pathways, thus preventing tumor cells from activating es-cape mechanisms. An additional challenge, however, is the need to effi-ciently elicit the response in all diseased cells, particularly for cancer.When a small number of cells evade treatment, the disease will prog-ress. Improvements in the delivery will greatly diminish this concern,but appropriate choice of miRNA may also minimize this concern. Byselecting miRNAs that will increase extracellular signaling molecules,treated cells will influence cells within the vicinity by these paracrinefactors.

For cells to respond to a miRNA therapeutic, separation from the de-livery system and subcellular localization ofmiRNA tomiRISC complexeswill be necessary. The majority of the activity of miRNA is related tomiRISC formation. Much like has been discussed for siRNA [194], thefact that inactive and active RISC has been localized to the endosomemay have great advantage for delivery. The location of RISC and possibil-ity that activity is localized to endosomes [130,131,195] suggests that es-cape from the endosomemay be unnecessary, and if escape is necessary,the localization to the endosomal membrane may increase activity. Thisis counter to the idea that P-bodies are the primary locations for miRISCactivitywhere P-body localizationwould be necessary for augmented in-tracellular activity. A thorough understanding of the biology behind theactivity of the therapeutic can be used to guide the design of the deliverysystem for optimal effectiveness to be achieved. As with any therapeuticapproach, the interaction between the therapeuticmolecule and deliverysystem is paramount to the activity of the final therapy.

The efficiency of miRNA-based therapeutic molecules rely largely onthe capacity of delivery agent to protect the oligonucleotides againstserumdegradation, accumulate to thediseased organor tissue via activetargeting, and elicit powerful therapeutic effect without causing un-wanted side effects. The delivery of miRNA is limited to parenteral orlocal injection due to the lack of enteral uptake of RNA although otherroutes are being examined [196]. Depending upon the disease, theamount of miRNA reaching a target organ (or organs) is limited. Thechemical challenges to stabilizing RNAs have been overcome bymodifi-cation and complexation, but the stability of the RNAs does not improvecellular uptake and escape. The vast majority of delivered RNA neverwill enter a cell, let alone the target cells. Each of the systems describedovercomes someof the numerous challenges, butmuch improvement ispossible.

5. Conclusion

Recent developments in the understanding of miRNA have sug-gested their therapeutic potential. But, the understanding of the biolog-ical genesis and activity will be the key to translation of this promisingtherapeutic class of molecules. Many of the challenges of deliveringmiRNA are similar, if not equivalent to siRNA and DNA, but some differ-ences exist. The primary differences exist in the specificity of mRNAsthat are repressed and the possibility to have promotion of expressiondirectly through the miRNA. Significant progress has been made inconfirming the ability to deliver miRNA, but substantial improvementswill be necessary. Delivery systemswill continue to be the key to bring-ing miRNA to the clinic.

Acknowledgments

This review and our original miRNA research was conducted in a fa-cility constructed with support from Research Facilities ImprovementProgram Grant (RR15482) from the National Centre for Research

972 Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

Resources (NCRR) of the National Institutes of Health (NIH). Our origi-nal miRNA delivery research has been funded, in part, by the Universityof Illinois at Chicago Center for Clinical and Translational Science (CCTS)award which is also supported by the NCRR (TR000050, RAG).

References

[1] B. Yan, Z.H. Wang, J.T. Guo, The research strategies for probing the function of longnoncoding RNAs, Genomics 99 (2012) 76–80.

[2] M. Wery, M. Kwapisz, A. Morillon, Noncoding RNAs in gene regulation, WileyInterdiscip. Rev. Syst. Biol. Med. 3 (2011) 728–738.

[3] S. Choudhuri, Small noncoding RNAs: biogenesis, function, and emerging signifi-cance in toxicology, J. Biochem. Mol. Toxicol. 24 (2010) 195–216.

[4] R. Robinson, RNAi therapeutics: how likely, how soon? PLoS Biol. 2 (2004) E28.[5] R.C. Lee, R.L. Feinbaum, V. Ambros, The C. elegans heterochronic gene lin-4 encodes

small RNAs with antisense complementarity to lin-14, Cell 75 (1993) 843–854.[6] B. Wightman, I. Ha, G. Ruvkun, Posttranscriptional regulation of the heterochronic

gene lin-14 by lin-4 mediates temporal pattern formation in C. elegans, Cell 75(1993) 855–862.

[7] A.E. Pasquinelli, B.J. Reinhart, et al., Conservation of the sequence and temporal ex-pression of let-7 heterochronic regulatory RNA, Nature 408 (2000) 86–89.

[8] B.J. Reinhart, F.J. Slack, et al., The 21-nucleotide let-7 RNA regulates developmentaltiming in Caenorhabditis elegans, Nature 403 (2000) 901–906.

[9] G.A. Calin, C.D. Dumitru, et al., Frequent deletions and down-regulation of micro-RNA genes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia, Proc.Natl. Acad. Sci. U. S. A. 99 (2002) 15524–15529.

[10] J. Li, Z. Zhang, miRNA regulatory variation in human evolution, Trends Genet. 29(2013) 116–124.

[11] V. Ambros, The functions of animal microRNAs, Nature 431 (2004) 350–355.[12] T. Braun, M. Gautel, Transcriptional mechanisms regulating skeletal muscle differ-

entiation, growth and homeostasis, Nat. Rev. Mol. Cell Biol. 12 (2011) 349–361.[13] R.T. Lima, S. Busacca, et al., MicroRNA regulation of core apoptosis pathways in can-

cer, Eur. J. Cancer 47 (2011) 163–174.[14] D.P. Bartel, MicroRNAs: genomics, biogenesis, mechanism, and function, Cell 116

(2004) 281–297.[15] L. He, G.J. Hannon, MicroRNAs: small RNAs with a big role in gene regulation, Nat.

Rev. Genet. 5 (2004) 522–531.[16] M.T. Bohnsack, K. Czaplinski, D. Gorlich, Exportin 5 is a RanGTP-dependent

dsRNA-binding protein that mediates nuclear export of pre-miRNAs, RNA-A publi-cation of the RNA society 10, 2004. 185–191.

[17] R. Yi, Y. Qin, et al., Exportin-5 mediates the nuclear export of pre-microRNAs andshort hairpin RNAs, Genes Dev. 17 (2003) 3011–3016.

[18] Y. Lee, C. Ahn, et al., The nuclear RNase III Drosha initiates microRNA processing,Nature 425 (2003) 415–419.

[19] A.M. Denli, B.B.J. Tops, et al., Processing of primary microRNAs by the microproces-sor complex, Nature 432 (2004) 231–235.

[20] L.P. Lim, N.C. Lau, et al., Microarray analysis shows that somemicroRNAs downreg-ulate large numbers of target mRNAs, Nature 433 (2005) 769–773.

[21] P.H. Olsen, V. Ambros, The lin-4 regulatory RNA controls developmental timing inCaenorhabditis elegans by blocking LIN-14 protein synthesis after the initiation oftranslation, Dev. Biol. 216 (1999) 671–680.

[22] J.G. Doench, C.P. Petersen, P.A. Sharp, siRNAs can function as miRNAs, Genes Dev.17 (2003) 438–442.

[23] M.S. Ebert, J.R. Neilson, P.A. Sharp, MicroRNA sponges: competitive inhibitors ofsmall RNAs in mammalian cells, Nat. Methods 4 (2007) 721–726.

[24] S. Vasudevan, Y. Tong, J.A. Steitz, Switching from repression to activation:microRNAs can up-regulate translation, Science 318 (2007) 1931–1934.

[25] H. Guo, N.T. Ingolia, et al., Mammalian microRNAs predominantly act to decreasetarget mRNA levels, Nature 466 (2010) 835–840.

[26] A. Gallo, M. Tandon, et al., The majority of microRNAs detectable in serum and sa-liva is concentrated in exosomes, PLoS One 7 (2012) e30679.

[27] L.C. Li, S.T. Okino, et al., Small dsRNAs induce transcriptional activation in humancells, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 17337–17342.

[28] X.Wang, X. Song, et al., The long arm of long noncoding RNAs: roles as sensors reg-ulating gene transcriptional programs, Cold Spring Harb. Perspect. Biol. 3 (2011)a003756.

[29] C. Holz-Schietinger, N.O. Reich, RNAmodulation of the humanDNAmethyltransferase3A, Nucleic Acids Res. 40 (2012) 8550–8557.

[30] E. Bernstein, C.D. Allis, RNA meets chromatin, Genes Dev. 19 (2005) 1635–1655.[31] C.D. Jeffries, H.M. Fried, D.O. Perkins, Nuclear and cytoplasmic localization of neural

stem cell microRNAs, RNA 17 (2011) 675–686.[32] D.H. Kim, P. Saetrom, et al., MicroRNA-directed transcriptional gene silencing in

mammalian cells, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 16230–16235.[33] S.T. Younger, D.R. Corey, Transcriptional gene silencing in mammalian cells

by miRNA mimics that target gene promoters, Nucleic Acids Res. 39 (2011)5682–5691.

[34] J. Tomikawa, H. Shimokawa, et al., Single-stranded noncoding RNAs mediate localepigenetic alterations at gene promoters in rat cell lines, J. Biol. Chem. 286 (2011)34788–34799.

[35] C.L. Jopling, M. Yi, et al., Modulation of hepatitis C virus RNA abundance by aliver-specific microRNA, Science 309 (2005) 1577–1581.

[36] A. Dharap, V.P. Nakka, R. Vemuganti, microRNAs in ischemic brain: the fine-tuningspecialists and novel therapeutic targets translational stroke research, in: P.A.Lapchak, J.H. Zhang (Eds.), Springer, US, 2012, pp. 335–352.

[37] J. Stanczyk, D.A.L. Pedrioli, et al., Altered expression of microRNA in synovial fibro-blasts and synovial tissue in rheumatoid arthritis, Arthritis Rheum. 58 (2008)1001–1009.

[38] J. Lu, G. Getz, et al., MicroRNA expression profiles classify human cancers, Nature435 (2005) 834–838.

[39] M. Tatsuguchi, H.Y. Seok, et al., Expression of microRNAs is dynamicallyregulated during cardiomyocyte hypertrophy, J. Mol. Cell. Cardiol. 42 (2007)1137–1141.

[40] K. Jeyaseelan, W.B. Herath, A. Armugam, MicroRNAs as therapeutic targets inhuman diseases, Expert Opin. Ther. Targets 11 (2007) 1119–1129.

[41] Y. Murakami, T. Yasuda, et al., Comprehensive analysis of microRNA expressionpatterns in hepatocellular carcinoma and non-tumorous tissues, Oncogene 25(2006) 2537–2545.

[42] P.T. Nelson, D.A. Baldwin, et al., Microarray-based, high-throughput gene expres-sion profiling of microRNAs, Nat. Methods 1 (2004) 155–161.

[43] K.A. Lennox, M.A. Behlke, Chemical modification and design of anti-miRNA oligo-nucleotides, Gene Ther. 18 (2011) 1111–1120.

[44] K.A. Lennox, M.A. Behlke, A direct comparison of anti-microRNA oligonucleotidepotency, Pharm. Res. 27 (2010) 1788–1799.

[45] F. Czauderna, M. Fechtner, et al., Structural variations and stabilising modifi-cations of synthetic siRNAs in mammalian cells, Nucleic Acids Res. 31 (2003)2705–2716.

[46] P.D. Cook, Medicinal chemistry of antisense oligonucleotides—future opportunities,Anticancer Drug Des. 6 (1991) 585–607.

[47] J. Krutzfeldt, N. Rajewsky, et al., Silencing of microRNAs in vivo with ‘antagomirs’,Nature 438 (2005) 685–689.

[48] J. Krutzfeldt, S. Kuwajima, et al., Specificity, duplex degradation and subcellular lo-calization of antagomirs, Nucleic Acids Res. 35 (2007) 2885–2892.

[49] S.T. Crooke, M.J. Graham, et al., Pharmacokinetic properties of several novel oligo-nucleotide analogs in mice, J. Pharmacol. Exp. Ther. 277 (1996) 923–937.

[50] B.H. Yoo, E. Bochkareva, et al., 2′-O-methyl-modified phosphorothioate antisenseoligonucleotides have reduced non-specific effects in vitro, Nucleic Acids Res. 32(2004) 2008–2016.

[51] C. Wahlestedt, P. Salmi, et al., Potent and nontoxic antisense oligonucleotides con-taining locked nucleic acids, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 5633–5638.

[52] B. Hyrup, P.E. Nielsen, Peptide nucleic acids (PNA): synthesis, properties and po-tential applications, Bioorg. Med. Chem. 4 (1996) 5–23.

[53] P.S. Pallan, E.M. Greene, et al., Unexpected origins of the enhanced pairing affinityof 2′-fluoro-modified RNA, Nucleic Acids Res. 39 (2011) 3482–3495.

[54] S. Akhtar, M.D. Hughes, et al., The delivery of antisense therapeutics, Adv. DrugDeliv. Rev. 44 (2000) 3–21.

[55] M.A. Behlke, Chemical modification of siRNAs for in vivo use, Oligonucleotides(2008) 18.

[56] M. Manoharan, RNA interference and chemically modified small interfering RNAs,Curr. Opin. Chem. Biol. 8 (2004) 570–579.

[57] D.R. Corey, Chemical modification: the key to clinical application of RNA interfer-ence? J. Clin. Invest. 117 (2007) 3615–3622.

[58] J.M. Campbell, T.A. Bacon, E. Wickstrom, Oligodeoxynucleoside phosphorothioatestability in subcellular extracts, culture media, sera and cerebrospinal fluid,J. Biochem. Biophys. Methods 20 (1990) 259–267.

[59] S.D. Patil, D.G. Rhodes, Influence of divalent cations on the conformation ofphosphorothioate oligodeoxynucleotides: a circular dichroism study, NucleicAcids Res. 28 (2000) 2439–2445.

[60] T.P. Prakash, A.M. Kawasaki, et al., Comparing in vitro and in vivo activity of2′-O-[2-(methylamino)-2-oxoethyl]- and 2′-O-methoxyethyl-modified antisenseoligonucleotides, J. Med. Chem. 51 (2008) 2766–2776.

[61] J. Elmen, M. Lindow, et al., LNA-mediated microRNA silencing in non-human pri-mates, Nature 452 (2008) 896-U810.

[62] R.E. Lanford, E.S. Hildebrandt-Eriksen, et al., Therapeutic silencing of microRNA-122in primates with chronic hepatitis C virus infection, Science 327 (2010) 198–201.

[63] H.L. Janssen, H.W. Reesink, et al., Treatment of HCV Infection by targetingMicroRNA, N. Engl. J. Med. 368 (2013) 1685–1694.

[64] M.M. Fabani, C. Abreu-Goodger, et al., Efficient inhibition of miR-155 functionin vivo by peptide nucleic acids, Nucleic Acids Res. 38 (2010) 4466–4475.

[65] M.M. Fabani, M.J. Gait, miR-122 targeting with LNA/2'-O-methyl oligonucleotidemixmers, peptide nucleic acids (PNA), and PNA, peptide conjugates, RNA, 14,2008. 336–346.

[66] R. Kole, A.R. Krainer, S. Altman, RNA therapeutics: beyond RNA interference andantisense oligonucleotides, Nat. Rev. Drug Discov. 11 (2012) 125–140.

[67] R.L. Juliano, X. Ming, O. Nakagawa, Cellular uptake and intracellular trafficking ofantisense and siRNA oligonucleotides, Bioconjug. Chem. 23 (2012) 147–157.

[68] T.B. Hansen, T.I. Jensen, et al., Natural RNA circles function as efficient microRNAsponges, Nature 495 (2013) 384–388.

[69] L. Ma, J. Young, et al., miR-9, a MYC/MYCN-activated microRNA, regulatesE-cadherin and cancer metastasis, Nat. Cell Biol. 12 (2010) 247–256.

[70] R. Garzon, G. Marcucci, C.M. Croce, Targeting microRNAs in cancer: rationale, strat-egies and challenges, Nat. Rev. Drug Discov. 9 (2010) 775–789.

[71] W.Y. Choi, A.J. Giraldez, A.F. Schier, Target protectors reveal dampening andbalancing of nodal agonist and antagonist bymiR-430, Science 318 (2007) 271–274.

[72] A.G. Bader, D. Brown, et al., Developing therapeutic microRNAs for cancer, GeneTher. 18 (2011) 1121–1126.

[73] D. Li, X. Liu, et al., MicroRNA-99a inhibits hepatocellular carcinoma growth andcorrelates with prognosis of patients with hepatocellular carcinoma, J. Biol.Chem. 286 (2011) 36677–36685.

[74] K. Terasawa, K. Shimizu, G. Tsujimoto, Synthetic pre-miRNA-based shRNA as po-tent RNAi triggers, J. Nucleic Acids 2011 (2011) 131579.

973Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

[75] K.A. Whitehead, R. Langer, D.G. Anderson, Knocking down barriers: advances insiRNA delivery, Nat. Rev. Drug Discov. 8 (2009) 129–138.

[76] W.J. Kim, S.W. Kim, Efficient siRNA delivery with non-viral polymeric vehicles,Pharm. Res. 26 (2009) 657–666.

[77] N. Khatri, M. Rathi, et al., In vivo delivery aspects of miRNA, shRNA and siRNA, Crit.Rev. Ther. Drug Carrier Syst. 29 (2012) 487–527.

[78] D.M. Pereira, P.M. Rodrigues, et al., Delivering the promise of miRNA cancer thera-peutics, Drug Discov. Today 18 (2013) 282–289.

[79] M.P. Czech, M. Aouadi, G.J. Tesz, RNAi-based therapeutic strategies for metabolicdisease, Nat. Rev. Endocrinol. 7 (2011) 473–484.

[80] T.G. Park, J.H. Jeong, S.W. Kim, Current status of polymeric gene delivery systems,Adv. Drug Deliv. Rev. 58 (2006) 467–486.

[81] D. Grimm, K.L. Streetz, et al., Fatality in mice due to oversaturation of cellularmicroRNA/short hairpin RNA pathways, Nature 441 (2006) 537–541.

[82] T.J. Harris, J.J. Green, et al., Tissue-specific gene delivery via nanoparticle coating,Biomaterials 31 (2010) 998–1006.

[83] K.C. Wood, S.M. Azarin, et al., Tumor-targeted gene delivery using molecularlyengineered hybrid polymers functionalized with a tumor-homing peptide,Bioconjug. Chem. 19 (2008) 403–405.

[84] H. Lv, S. Zhang, et al., Toxicity of cationic lipids and cationic polymers in gene de-livery, J. Control. Release 114 (2006) 100–109.

[85] Y.C. Tseng, S. Mozumdar, L. Huang, Lipid-based systemic delivery of siRNA, Adv.Drug Deliv. Rev. 61 (2009) 721–731.

[86] J. Elmen, M. Lindow, et al., Antagonism of microRNA-122 in mice by systemicallyadministered LNA-antimiR leads to up-regulation of a large set of predicted targetmRNAs in the liver, Nucleic Acids Res. 36 (2008) 1153–1162.

[87] N. Liu, C. Zhou, et al., Reversal of paclitaxel resistance in epithelial ovariancarcinoma cells by an MUC1 aptamer-let-7i chimera, Cancer Invest. 30(2012) 577–582.

[88] Y. Wu,M. Crawford, et al., MicroRNA delivery by cationic lipoplexes for lung cancertherapy, Mol. Pharm. 8 (2011) 1381–1389.

[89] V.J. Craig, A. Tzankov, et al., Systemic microRNA-34a delivery induces apoptosisand abrogates growth of diffuse large B-cell lymphoma in vivo, Leukemia 26(2012) 2421–2424.

[90] P. Trang, J.F. Wiggins, et al., Systemic delivery of tumor suppressor microRNAmimics using a neutral lipid emulsion inhibits lung tumors in mice, Mol. Ther. 19(2011) 1116–1122.

[91] Y. Akao, Y. Nakagawa, et al., Role of anti-oncomirs miR-143 and -145 in human co-lorectal tumors, Cancer Gene Ther. 17 (2010) 398–408.

[92] A. Akinc, A. Zumbuehl, et al., A combinatorial library of lipid-like materials for de-livery of RNAi therapeutics, Nat. Biotechnol. 26 (2008) 561–569.

[93] L. Piao, M. Zhang, et al., Lipid-based nanoparticle delivery of pre-miR-107 inhibitsthe tumorigenicity of head and neck squamous cell carcinoma, Mol. Ther. 20(2012) 1261–1269.

[94] Y. Chen, X. Zhu, et al., Nanoparticles modified with tumor-targeting scFv deliversiRNA and miRNA for cancer therapy, Mol. Ther. 18 (2010) 1650–1656.

[95] X.Q. Liu, W.J. Song, et al., Targeted delivery of antisense inhibitor of miRNA forantiangiogenesis therapy using cRGD-functionalized nanoparticles, Mol. Pharm. 8(2011) 250–259.

[96] J. Su, H. Baigude, et al., Silencing microRNA by interfering nanoparticles in mice,Nucleic Acids Res. 39 (2011) e38.

[97] K. Rai, N. Takigawa, et al., Liposomal delivery of MicroRNA-7-expressing plasmidovercomes epidermal growth factor receptor tyrosine kinase inhibitor-resistancein lung cancer cells, Mol. Cancer Ther. 10 (2011) 1720–1727.

[98] S. Shi, L. Han, et al., Systemic delivery of microRNA-34a for cancer stem cell thera-py, Angew. Chem. 52 (2013) 3901–3905.

[99] A.F. Ibrahim, U. Weirauch, et al., MicroRNA replacement therapy for miR-145 andmiR-33a is efficacious in a model of colon carcinoma, Cancer Res. 71 (2011)5214–5224.

[100] G.Y. Chiou, J.Y. Cherng, et al., Cationic polyurethanes-short branch PEI-mediateddelivery of Mir145 inhibited epithelial-mesenchymal transdifferentiation andcancer stem-like properties and in lung adenocarcinoma, J. Control. Release 159(2012) 240–250.

[101] W. Hwang do, S. Son, et al., A brain-targeted rabies virus glycoprotein-disulfidelinked PEI nanocarrier for delivery of neurogenic microRNA, Biomaterials 32(2011) 4968–4975.

[102] Y. Ren, X. Zhou, et al., MicroRNA-21 inhibitor sensitizes human glioblastomacells U251 (PTEN-mutant) and LN229 (PTEN-wild type) to taxol, BMC Cancer 10(2010) 27.

[103] Y. Ren, C.S. Kang, et al., Co-delivery of as-miR-21 and 5-FU by poly(amidoamine)dendrimer attenuates human glioma cell growth in vitro, J. Biomater. Sci. Polym.Ed. 21 (2010) 303–314.

[104] C.J. Cheng, W.M. Saltzman, Polymer nanoparticle-mediated delivery of MicroRNAinhibition and alternative splicing, Mol. Pharm. 9 (2012) 1481–1488.

[105] I.A. Babar, C.J. Cheng, et al., Nanoparticle-based therapy in an in vivo microRNA-155(miR-155)-dependent mouse model of lymphoma, Proc. Natl. Acad. Sci. U. S. A.109 (2012) E1695–E1704.

[106] H. Tazawa, N. Tsuchiya, et al., Tumor-suppressivemiR-34a induces senescence-likegrowth arrest throughmodulation of the E2F pathway in human colon cancer cells,Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 15472–15477.

[107] H. Matsuyama, H.I. Suzuki, et al., miR-135b mediates NPM-ALK-driven oncogenic-ity and renders IL-17-producing immunophenotype to anaplastic large cell lym-phoma, Blood 118 (2011) 6881–6892.

[108] F. Takeshita, L. Patrawala, et al., Systemic delivery of synthetic microRNA-16 in-hibits the growth of metastatic prostate tumors via downregulation of multiplecell-cycle genes, Mol. Ther. J. Am. Soc. Gene Ther. 18 (2010) 181–187.

[109] S. Ohno, M. Takanashi, et al., Systemically injected exosomes targeted to EGFR de-liver antitumor microRNA to breast cancer cells, Mol. Ther. 21 (2013) 185–191.

[110] A. Tivnan, W.S. Orr, et al., Inhibition of neuroblastoma tumor growth by targeteddelivery ofmicroRNA-34ausing anti-disialogangliosideGD(2) coated nanoparticles,PLoS One 7 (2012) e38129.

[111] M.V. Yigit, S.K. Ghosh, et al., Context-dependent differences in miR-10b breast on-cogenesis can be targeted for the prevention and arrest of lymph node metastasis,Oncogene 32 (2013) 1530–1538.

[112] O.P. Medina, Y. Zhu, K. Kairemo, Targeted liposomal drug delivery in cancer, Curr.Pharm. Des. 10 (2004) 2981–2989.

[113] H. Ando, A. Okamoto, et al., Development of a miR-92a delivery system foranti-angiogenesis-based cancer therapy, J. Gene Med. 15 (2013) 20–27.

[114] H. Baigude, J. McCarroll, et al., Design and creation of new nanomaterials for ther-apeutic RNAi, ACS Chem. Biol. 2 (2007) 237–241.

[115] O. Boussif, F. Lezoualc'h, et al., A versatile vector for gene and oligonucleotide trans-fer into cells in culture and in vivo: polyethylenimine, Proc. Natl. Acad. Sci. U. S. A.92 (1995).

[116] A. de Fougerolles, H.P. Vornlocher, et al., Interfering with disease: a progress reporton siRNA-based therapeutics, Nat. Rev. Drug Discov. 6 (2007) 443–453.

[117] L. BioCancell, Phase 2b, trial of intravesical DTA-H19/PEI in patients withintermediate-risk superficial bladder cancer, ClinicalTrials.gov [Internet], NationalLibrary of Medicine (US), Bethesda, MD, 2008, ([Cited July 18, 2013]. NLM Identi-fier: NCT00595088).

[118] BioCancell, Ltd., Pilot study of BC-819/PEI and BCG in patients with recurrentsuperficial transitional cell bladder carcinoma, ClinicalTrials.gov [Internet], Nation-al Library of Medicine (US), Bethesda, MD, 2013, ([Cited July 18, 2013]. NLMIdentifier: NCT01878188).

[119] Y.P. Yang, Y. Chien, et al., Inhibition of cancer stem cell-like properties and reducedchemoradioresistance of glioblastoma using microRNA145 with cationicpolyurethane-short branch PEI, Biomaterials 33 (2012) 1462–1476.

[120] P. Kumar, H. Wu, et al., Transvascular delivery of small interfering RNA to the cen-tral nervous system, Nature 448 (2007) 39–43.

[121] N.Maiorano, A.Mallamaci, Promotion of embryonic cortico-cerebral neuronogenesisby miR-124, Neural Dev. 4 (2009) 40.

[122] Q.L. Hu, Q.Y. Jiang, et al., Cationic microRNA-delivering nanovectors with bifunc-tional peptides for efficient treatment of PANC-1 xenograft model, Biomaterials34 (2013) 2265–2276.

[123] T. Dutta, N.K. Jain, et al., Dendrimer nanocarriers as versatile vectors in gene deliv-ery, Nanomedicine 6 (2010) 25–34.

[124] S. Hong, A.U. Bielinska, et al., Interaction of poly(amidoamine) dendrimers withsupported lipid bilayers and cells: Hole formation and the relation to transport,Bioconjug. Chem. 15 (2004) 774–782.

[125] R.K. Kulkarni, E.G. Moore, et al., Biodegradable poly(lactic acid) polymers,J. Biomed. Mater. Res. 5 (1971) 169–181.

[126] J.S. Blum, W.M. Saltzman, High loading efficiency and tunable release of plasmidDNA encapsulated in submicron particles fabricated from PLGA conjugated withpoly-L-lysine, J. Control. Release 129 (2008) 66–72.

[127] K. Ramanlal Chaudhari, A. Kumar, et al., Bone metastasis targeting: a novel ap-proach to reach bone using Zoledronate anchored PLGA nanoparticle as carrier sys-tem loaded with Docetaxel, J. Control. Release 158 (2012) 470–478.

[128] I.F. Uchegbu, Pharmaceutical nanotechnology: polymeric vesicles for drug andgene delivery, Expert Opin. Drug Deliv. 3 (2006) 629–640.

[129] H. Maeda, The enhanced permeability and retention (EPR) effect in tumor vascula-ture: the key role of tumor-selective macromolecular drug targeting, Adv. Enzym.Regul. 41 (2001) 189–207.

[130] D. Gibbings, O. Voinnet, Control of RNA silencing and localization by endolysosomes,Trends Cell Biol. 20 (2010) 491–501.

[131] D.J. Gibbings, C. Ciaudo, et al., Multivesicular bodies associate with components ofmiRNA effector complexes and modulate miRNA activity, Nat. Cell Biol. 11 (2009)1143–1149.

[132] R.S. Gomes, R.P. Neves, et al., Efficient pro-survival/angiogenic miRNA delivery byan MRI-detectable nanomaterial, ACS Nano 7 (2013) 3362–3372.

[133] J.M. Dang, K.W. Leong, Natural polymers for gene delivery and tissue engineering,Adv. Drug Deliv. Rev. 58 (2006) 487–499.

[134] R.I. Mahat, O.D. Monera, et al., Peptide-based gene delivery, Curr. Opin. Mol. Ther. 1(1999) 226–243.

[135] B.D. Ratner, A.S. Hoffman, F.J. Schoen, Biomaterials science: an introduction to ma-terials in medicine, Academic Press, 1996.

[136] K.H. Stenzel, T. Miyata, A.L. Rubin, Collagen as a biomaterial, Annu. Rev. Biophys.Bioeng. 3 (1974) 231–253.

[137] Y. Minakuchi, F. Takeshita, et al., Atelocollagen-mediated synthetic small interfer-ing RNA delivery for effective gene silencing in vitro and in vivo, Nucleic AcidsRes. 32 (2004) e109.

[138] J.S. Suh, J.Y. Lee, et al., Peptide-mediated intracellular delivery ofmiRNA-29b for os-teogenic stem cell differentiation, Biomaterials 34 (2013) 4347–4359.

[139] E. Crew, S. Rahman, et al., MicroRNA conjugated gold nanoparticles and cell trans-fection, Anal. Chem. 84 (2012) 26–29.

[140] R. Ghosh, L.C. Singh, et al., A gold nanoparticle platform for the delivery of function-al microRNAs into cancer cells, Biomaterials 34 (2013) 807–816.

[141] A. Bitar, N.M. Ahmad, et al., Silica-based nanoparticles for biomedical applications,Drug Discov. Today 17 (2012) 1147–1154.

[142] A.S. Hoffman, The origins and evolution of controlled drug delivery systems, J. Con-trol. Release 132 (2008) 153–163.

[143] A.M. Krieg, Is RNAi dead? Mol. Ther. 19 (2011) 1001–1002.[144] H. Ledford, Drug giants turn their backs on RNA interference, Nature 468 (2010)

487.

974 Y. Zhang et al. / Journal of Controlled Release 172 (2013) 962–974

[145] Santaris Pharma A/S, Miravirsen in combination with telaprevir and ribavirin innull responder to pegylated-interferon alpha plus ribavirin subjects with chronichepatitis C virus infection, ClinicalTrias.gov [Internet], National Library of Medicine(US), Bethesda, MD, 2013, ([Cited July 19, 2013]. NLM Identifier: NCT01872936).

[146] Santaris Pharma A/S, Drug interaction study to assess the effect of co-administeredmiravirsen and telaprevir in healthy subjects, ClinicalTrias.gov [Internet], NationalLibrary of Medicine (US), Bethesda, MD, 2012, ([Cited July 19, 2013]. NLM Identi-fier: NCT01646489).

[147] Santaris Pharma A/S, Miravirsen study in null responder to pegylated interferonalpha plus ribavirin subjects with chronic hepatitis C, ClinicalTrias.gov [Internet],National Library of Medicine (US), Bethesda, MD, 2012, ([Cited July 19, 2013].NLM Identifier: NCT01727934).

[148] Santaris Pharma A/S, Multiple ascending dose study of miravirsen in treatment-naïve chronic hepatitis c subjects, ClinicalTrias.gov [Internet], National Library ofMedicine (US), Bethesda, MD, 2010–2012, ([Cited July 19, 2013]. NLM Identifier:NCT01200420).

[149] Santaris Pharma A/S, SPC3649 multiple dose study in healthy volunteers, NationalLibrary of Medicine (US), Bethesda, MD, 2009–2011, ([Cited July 19, 2013]. NLMIdentifier: NCT00979927).

[150] Santaris Pharma A/S, Safety study of SPC3649 in healthy men, ClinicalTrias.gov[Internet], National Library of Medicine (US), Bethesda, MD, 2008–2009, ([CitedJuly 18, 2013]. NLM Identifier: NCT00688012).

[151] J. Guo, J.C. Evans, C.M. O'Driscoll, Delivering RNAi therapeutics with non-viraltechnology: a promising strategy for prostate cancer? Trends Mol. Med. 19(2013) 250–261.

[152] Mirna Therapeutics, Inc., A phase I study of MRX34 given intravenously in patientswith unresectable primary liver cancer or metastatic cancer with liver involve-ment, National Library of Medicine (US), Bethesda, MD, 2013, ([Cited July 18,2013]. NLM Identifier: NCT01829971).

[153] A.C. Eifler, C.S. Thaxton, Nanoparticle therapeutics: FDA approval, clinical trials,regulatory pathways, and case study, Methods Mol. Biol. 726 (2011) 325–338.

[154] C. Scholz, E. Wagner, Therapeutic plasmid DNA versus siRNA delivery: commonand different tasks for synthetic carriers, J. Control. Release 161 (2012) 554–565.

[155] D. Haussecker, The business of RNAi therapeutics in 2012, Mol. Ther. Nucleic Acids1 (2012) e8.

[156] E. Ruoslahti, S.N. Bhatia, M.J. Sailor, Targeting of drugs and nanoparticles to tumors,J. Cell Biol. 188 (2010) 759–768.

[157] R.Oberbauer, G.F. Schreiner, T.W.Meyer, Renal uptake of an18-merphosphorothioateoligonucleotide, Kidney Int. 48 (1995) 1226–1232.

[158] F.M. Veronese, Peptide and protein PEGylation: a review of problems and solu-tions, Biomaterials 22 (2001) 405–417.

[159] F.M. Veronese, G. Pasut, PEGylation, successful approach to drug delivery, DrugDiscov. Today 10 (2005) 1451–1458.

[160] S. Friman, B. Egestad, et al., Hepatic excretion and metabolism of polyethylene gly-cols and mannitol in the cat, J. Hepatol. 17 (1993) 48–55.

[161] M.A. Dobrovolskaia, P. Aggarwal, et al., Preclinical studies to understand nanopar-ticle interaction with the immune system and its potential effects on nanoparticlebiodistribution, Mol. Pharm. 5 (2008) 487–495.

[162] M.A. Dobrovolskaia, S.E. McNeil, Immunological properties of engineered nano-materials, Nat. Nanotechnol. 2 (2007) 469–478.

[163] J. Rao, Shedding light on tumors using nanoparticles, ACS Nano 2 (2008)1984–1986.

[164] S.M. Moghimi, A.C. Hunter, J.C. Murray, Long-circulating and target-specificnanoparticles: theory to practice, Pharmacol. Rev. 53 (2001) 283–318.

[165] R.K. Jain, Barriers to drug-delivery in solid tumors, Sci. Am. 271 (1994) 58–65.[166] R.K. Jain, Delivery of molecular medicine to solid tumors: lessons from in vivo im-

aging of gene expression and function, J. Control. Release 74 (2001) 7–25.[167] R.K. Jain, T. Stylianopoulos, Delivering nanomedicine to solid tumors, Nat. Rev. Clin.

Oncol. 7 (2010) 653–664.[168] J.B. Bramsen, M.B. Laursen, et al., A large-scale chemical modification screen iden-

tifies design rules to generate siRNAs with high activity, high stability and low tox-icity, Nucleic Acids Res. 37 (2009) 2867–2881.

[169] Y.H. Bae, K. Park, Targeted drug delivery to tumors: myths, reality and possibility,J. Control. Release 153 (2011) 198–205.

[170] A.A. Khan, D. Betel, et al., Transfection of small RNAs globally perturbs gene regu-lation by endogenous microRNAs, Nat. Biotechnol. 27 (2009) 549–555.

[171] Y. Murakami, H. Toyoda, et al., The progression of liver fibrosis is relatedwith over-expression of the miR-199 and 200 families, PLoS One 6 (2011) e16081.

[172] E.M. Small, R.J. Frost, E.N. Olson, MicroRNAs add a new dimension to cardiovasculardisease, Circulation 121 (2010) 1022–1032.

[173] S. Das, K. Bryan, et al., Modulation of neuroblastoma disease pathogenesis by an ex-tensive network of epigenetically regulated microRNAs, Oncogene 32 (2013)2927–2936.

[174] A.J. Bridge, S. Pebernard, et al., Induction of an interferon response by RNAi vectorsin mammalian cells, Nat. Genet. 34 (2003) 263–264.

[175] V. Hornung, M. Guenthner-Biller, et al., Sequence-specific potent induction ofIFN-alpha by short interfering RNA in plasmacytoid dendritic cells through TLR7,Nat. Med. 11 (2005) 263–270.

[176] A.D. Judge, V. Sood, et al., Sequence-dependent stimulation of the mammalian in-nate immune response by synthetic siRNA, Nat. Biotechnol. 23 (2005) 457–462.

[177] C.A. Sledz, M. Holko, et al., Activation of the interferon system by short-interferingRNAs, Nat. Cell Biol. 5 (2003) 834–839.

[178] X. He, M. Pool, et al., Knockdown of polypyrimidine tract-binding protein sup-presses ovarian tumor cell growth and invasiveness in vitro, Oncogene 26 (2007)4961–4968.

[179] S.T. Ong, F. Li, et al., Hybrid cytomegalovirus enhancer-h1 promoter-based plasmidand baculovirus vectors mediate effective RNA interference, Hum. Gene Ther. 16(2005) 1404–1412.

[180] S. Samakoglu, L. Lisowski, et al., A genetic strategy to treat sickle cell anemia bycoregulating globin transgene expression and RNA interference, Nat. Biotechnol.24 (2006) 89–94.

[181] S. Xiang, J. Fruehauf, C.J. Li, Short hairpin RNA-expressing bacteria elicit RNA inter-ference in mammals, Nat. Biotechnol. 24 (2006) 697–702.

[182] A.M. Krichevsky, K.S. King, et al., A microRNA array reveals extensive regulation ofmicroRNAs during brain development, RNA-a publication of the rna society 9,2003. 1274–1281.

[183] S.C. Ling, J. Zhou, et al., The recent updates of therapeutic approaches against A betafor the treatment of Alzheimer's disease, Anat. Rec. 294 (2011) 1307–1318.

[184] K.R. Cordes, N.T. Sheehy, et al., miR-145 and miR-143 regulate smooth muscle cellfate and plasticity, Nature 460 (2009) 705–710.

[185] D.C. Altieri, Survivin, cancer networks and pathway-directed drug discovery, Nat.Rev. Cancer 8 (2008) 61–70.

[186] Z. Wang, The guideline of the design and validation of MiRNA mimics, MethodsMol. Biol. 676 (2011) 211–223.

[187] C.R. Sibley, Y. Seow,M.J.Wood, Novel RNA-based strategies for therapeutic gene si-lencing, Mol. Ther. 18 (2010) 466–476.

[188] B.-B.S. Zhou, H. Zhang, et al., Tumour-initiating cells: challenges and opportunitiesfor anticancer drug discovery, Nat. Rev. Drug Discov. 8 (2009) 806–823.

[189] W.C. Cho, Exploiting the therapeutic potential of microRNAs in human cancer,Expert Opin. Ther. Targets 16 (2012) 345–350.

[190] M. Sachdeva, Y.Y. Mo, MicroRNA-145 suppresses cell invasion and metastasis bydirectly targeting mucin 1, Cancer Res. 70 (2010) 378–387.

[191] M. Sachdeva, H. Wu, et al., MicroRNA-101-mediated Akt activation and estrogen-independent growth, Oncogene 30 (2011) 822–831.

[192] S. Zhu, M.L. Si, et al., MicroRNA-21 targets the tumor suppressor gene tropomyosin1 (TPM1), J. Biol. Chem. 282 (2007) 14328–14336.

[193] S. Zhu, H. Wu, et al., MicroRNA-21 targets tumor suppressor genes in invasion andmetastasis, Cell Res. 18 (2008) 350–359.

[194] J. Nguyen, F.C. Szoka, Nucleic acid delivery: the missing pieces of the puzzle? Acc.Chem. Res. 45 (2012) 1153–1162.

[195] Y.S. Lee, S. Pressman, et al., Silencing by small RNAs is linked to endosomal traffick-ing, Nat. Cell Biol. 11 (2009) 1150–1156.

[196] D.C. Forbes, N.A. Peppas, Oral delivery of small RNA and DNA, J. Control. Release162 (2012) 438–445.