Efficient and targeted delivery of siRNA in vivo

Authors

  • Min Suk Shim,

    1.  Department of Chemical Engineering and Materials Science, University of California, Irvine, CA, USA
    Search for more papers by this author
  • Young Jik Kwon

    1.  Department of Chemical Engineering and Materials Science, University of California, Irvine, CA, USA
    2.  Department of Pharmaceutical Sciences, University of California, Irvine, CA, USA
    3.  Department of Biomedical Engineering, University of California, Irvine, CA, USA
    Search for more papers by this author

Y. J. Kwon, Department of Pharmaceutical Sciences, 916 Engineering Tower, University of California, Irvine, CA 92697, USA
Fax: +1 949 824 2541
Tel: +1 949 824 8714
E-mail: kwonyj@uci.edu

Abstract

RNA interference (RNAi) has been regarded as a revolutionary tool for manipulating target biological processes as well as an emerging and promising therapeutic strategy. In contrast to the tangible and obvious effectiveness of RNAi in vitro, silencing target gene expression in vivo using small interfering RNA (siRNA) has been a very challenging task due to multiscale barriers, including rapid excretion, low stability in blood serum, nonspecific accumulation in tissues, poor cellular uptake and inefficient intracellular release. This minireview introduces major challenges in achieving efficient siRNA delivery in vivo and discusses recent advances in overcoming them using chemically modified siRNA, viral siRNA vectors and nonviral siRNA carriers. Enhanced specificity and efficiency of RNAi in vivo via selective accumulations in desired tissues, specific binding to target cells and facilitated intracellular trafficking are also commonly attempted utilizing targeting moieties, cell-penetrating peptides, fusogenic peptides and stimuli-responsive polymers. Overall, the crucial roles of the interdisciplinary approaches to optimizing RNAi in vivo, by efficiently and specifically delivering siRNA to target tissues and cells, are highlighted.

Abbreviations
ApoB

apolipoprotein B

CPP

cell-penetrating peptide

FA

folic acid

GFP

green fluorescent protein

HER-2

human epidermal growth factor 2

i.p.

intraperitoneal

i.t.

intratumoral

i.v.

intravenous

9R

nonamer arginine residues

RGD

Arg-Gly-Asp peptide

RNAi

RNA interference

siRNA

small interfering RNA

VEGF

vascular endothelial growth factor

Introduction

RNA interference (RNAi) is a highly conserved biological process among yeasts, worms, insects, plants and humans [1]. A single strand of exogenously introduced double-stranded small interfering RNA (siRNA; 20–30 nucleotides) guides an RNA-inducing silencing protein complex to degrade the mRNA with the matching sequence; thus, translation into the target proteins is silenced [2–4]. RNAi has been of great interest not only as a powerful research tool to suppress the expression of a target gene, but also as an emerging therapeutic strategy to silence disease genes [5]. Theoretically, siRNA can interfere with the translation of almost any mRNA, as long as the mRNA has a distinctive sequence, whereas the targets of traditional drugs are limited by types of cellular receptors and enzymes [6].

Cancer, viral infections, autoimmune diseases and neurodegenerative diseases have been explored as promising disease targets of RNAi [7,8]. Recent progress in clinical trials using siRNA to cure age-related macular degeneration (bevasiranib; Opko Health, Inc., Miami, FL, USA; phase III) and respiratory syncytial virus infection (ALN-RSV01; Alnylam, Cambridge, MA, USA; phase II) have demonstrated the therapeutic potential of RNAi [9]. Moreover, the first evidence of targeted in vivo gene silencing for human cancer therapy via systemic delivery of siRNA using transferrin-tagged, cyclodextrin-based polymeric nanoparticles (CALAA-01; Calando Pharmaceuticals, Pasadena, CA, USA; phase I) has been recently announced [10].

Despite quite efficient and reliable gene silencing in vitro, only limited RNAi has been achieved in vivo because of rapid enzymatic degradation in combination with poor cellular uptake of siRNA [11]. Therefore, novel delivery systems, which enable prolonged circulation of siRNA with resistance against enzymatic degradation, high accessibility to target cells via clinically feasible administration routes and optimized cytosolic release of siRNA after efficient cellular uptake, are indispensably required [12]. In this minireview, major factors in determining overall RNAi efficiency in vivo are introduced. Moreover, up-to-date progress in achieving efficient and targeted siRNA delivery in vivo, particularly by overcoming multiscale hurdles using novel siRNA carriers, is discussed.

Challenges in RNAi in vivo

Design and in vivo delivery of siRNA

There are multiple key considerations in order to achieve efficient RNAi in vivo by delivering exogenous siRNA. siRNA has to be designed to target hybridization-accessible regions within the target mRNA while avoiding unintended (off-target) effects [13–15], which is extensively reviewed in this series by Walton et al. [16]. In addition, siRNA can also induce adverse effects such as immune responses, as discussed by Samuel-Abraham & Leonard [17]. siRNA may induce interferon responses either through the double-stranded RNA-activated protein kinase PKR [18] or toll-like receptor 3 [19]. Therefore, a combination of computer algorithms and experimental validation should be employed to determine the optimized siRNA sequences that are complementary to target mRNA while inducing minimal immune responses [20].

Naked siRNA is relatively unstable in blood in its native form and is rapidly cleared from the body (i.e. short half-lives in vivo) via degradion by ribonucleases, rapid renal excretion and nonspecific uptake by the reticuloendothelial system [21]. The phosphorothioate backbone, or various 2′ positions in the sugar moiety of siRNA, is conventionally modified to enhance its stability and activity against nuclease degradation [22,23], without affecting gene silencing activity [24]. siRNA is an anionic macromolecule and does not readily enter cells by passive diffusion mechanisms. An appropriate siRNA delivery system enhances cellular uptake, protects its payload from enzymatic digestion and immune recognition, and improves the pharmacokinetics by avoiding excretion via the reticuloendothelial system and renal filtration (i.e. prolonged half-life in vivo) [25–27]. In addition, targeted delivery systems localize siRNA in the desired tissue, resulting in a reduction in the amount of siRNA required for efficient gene silencing in vivo, as well as minimized side effects. Therefore, the development of effective in vivo delivery systems is pivotal in overcoming the challenges in achieving efficient and targeted siRNA delivery in vivo. Major hurdles in siRNA delivery in vivo and various approaches to overcoming them are illustrated in Fig. 1.

Figure 1.

 Interdisciplinary approaches to achieving efficient and targeted RNAi in vivo by overcoming multiscale barriers in systemic siRNA delivery. Detailed design parameters of an ideal siRNA carrier are depicted in Fig. 2.

Local versus systemic delivery

The types of target tissues and cells dictate the optimum administration routes of local versus systemic delivery. For example, siRNA can be directly applied to the eye, skin or muscle via local delivery, whereas systemic siRNA delivery is the only way to reach metastatic and hematological cancer cells. Local delivery offers several advantages over systemic delivery, such as low effective doses, simple formulation (e.g. no targeting moieties), low risk of inducing systemic side effects and facilitated site-specific delivery [28]. Therefore, if applicable, local delivery is likely to be a more cost-efficient strategy for siRNA delivery in vivo than systemic administration. For example, initial clinical trials for RNAi-based treatment of age-related macular degeneration have exclusively used local injections of siRNA directly into the eye [10]. Other promising local routes include intranasal siRNA administration for pulmonary delivery [10,29–31] and direct injection into the central nervous system [10,32,33].

Alternatively, systemic delivery via intravenous (i.v.), intraperitoneal (i.p.) or oral administration is widely applicable when the target sites are not locally confined or not readily accessible. Metastatic tumors are especially amenable for systemic delivery compared with local administration. For example, human bcl-2 oncogene-targeting siRNA, which was complexed with cationic liposomes and i.v. injected, effectively inhibited tumor growth in a mouse liver metastasis model [34]. Another study showed that siRNA encapsulated in a lipid vesicle was able to impart efficient and persistent antiviral activity after being injected into a hepatitis B virus mouse model [35]. However, importantly, systemic siRNA delivery imposes several additional barriers in comparison with local delivery. siRNA should remain in its active form during circulation and be able to reach target tissues after passing through multiple barrier organs (e.g. liver, kidney and lymphoid organs).

Extracellular and intracellular barriers in siRNA delivery in vivo

Regardless of administration routes, the final destination of siRNA is the cytoplasm of the target cell, where it incorporates into RNA-inducing silencing protein complex and encounters target mRNAs. First, siRNA that survives in the plasma and is transported close to a target tissue must extravasate through the tight vascular endothelial junctions. It has been reported that microvascular transport of macromoelcules > 5 nm in diameter is significantly inhibited in normal tissues [36]. However, transport of macromolecules across the tumor endothelium is more efficient than that of normal endothelium because of its leaky and discontinuous vascular structures with poor lymphatic drainage. Thus, tumor endothelium allows the penetration of high molecular mass macromolecules (> 40 kDa), which is also referred to as ‘enhanced permeation and retention effect’ [37]. siRNA, in its native form or formulated in a delivery carrier, must then diffuse through the extracellular matrix, a dense network of fibrous protein and carbohydrates surrounding a cell [38], before accessing target cells. siRNA or its complex adheres preferably to target cells via receptor-mediated specific binding, followed by cellular uptake. Even after it is internalized by a cell, siRNA should be released from the endosome, while avoiding entrapment and degradation [39,40]. Because the condition in the endosome/lysosome is mildly acidic, facilitated cytosolic release of siRNA using acid-responsive delivery carriers has been a popular strategy to overcome this intracellular hurdle [41,42]. Fusogenic peptides which undergo acid-triggered conformational changes have also been shown to accelerate endosomal escape of nucleic acids [43,44]. Finally, siRNA delivered by a carrier should be decomplexed in the cytoplasm [45]. A broad range of novel materials that provide enhanced siRNA release have been developed (e.g. disulfide-based cationic polymers) [46]. Fig. 1 shows extracellular and intracellular barriers in siRNA delivery with various approaches to overcoming them.

Chemically modified siRNA for enhanced RNAi in vivo

Various molecular positions in siRNA have been chemically replaced or modified, mainly to resist enzymatic hydrolysis. For example, phosphodiester (PO4) linkages were replaced with phosphothioate (PS) at the 3′-end, and introducing O-methyl (2′-O-Me), fluoro (2′-F) group or methoxyethyl (2′-O-MOE) group greatly prolonged half-lives in plasma and enhanced RNAi efficiency in cultured cells [47–51]. In addition, efficiency enhancer molecules were conjugated to either the 5′- or 3′-end of the sense strand, without affecting the activity of the antisense strand [52]. There are some potential risks that chemically modifying siRNA may compromise RNAi efficiency. For example, boranophosphonate modification at the central position of the antisense strand of siRNA showed improved resistance to nuclease degradation, but simultaneously reduced RNAi activity [53]. In addition, non-natural molecules produced upon the degradation of a chemically modified siRNA may generate metabolites that might be unsafe or trigger unwanted effects. To date, cholesterol and aptamers are the most promising siRNA conjugates that have demonstrated efficient RNAi in vivo.

Cholesterol–siRNA conjugates

Improved pharmacokinetic and cellular uptake properties of cholesterol–siRNA conjugates silenced apolipoprotein B (ApoB) in mice via i.v. administration [22]. By contrast, ApoB siRNA unconjugated with cholesterol was unable to induce mRNA interference and was rapidly cleared. The mechanisms of improved distribution and cellular uptake of siRNA through cholesterol conjugation were demonstrated in a recent study; cholesterol–siRNA conjugates seem to incorporate into circulating lipoprotein particles (i.e. improved distribution in vivo) and are efficiently internalized by hepatocytes via receptor-mediated processes (i.e. efficient cellular uptake) [54]. Prebinding of cholesterol–siRNA conjugates to lipoparticles dramatically improved silencing efficacy in mice, and lipoparticle types affected cholesterol–siRNA conjugate distribution in various tissues [54]. Using a transgenic mouse model for Huntington’s disease, it was also demonstrated that a single intrastriatal injection of cholesterol–siRNA conjugates silenced a mutant Huntingtin gene, attenuating neuronal pathology as well as delaying the abnormal behavioral phenotype [55].

RNA aptamer–siRNA conjugates

RNA aptamers have been popularly used to selectively deliver siRNA in vivo to target tissues and cells, such as prostate cancer cells and tumor vascular endothelium overexpressing prostate-specific membrane antigen [56]. A key advantage of aptamer-mediated targeted delivery systems is that RNA aptamers can be facilely obtained by in vitro transcription reaction and, therefore, avoid contamination by cell or bacterial products. Promising in vitro and in vivo RNAi was obtained using siRNA that was directly linked with prostate-specific membrane antigen aptamers [57]. An aptamer-based delivery system has also been used to suppress HIV infection. Anti-gp120 RNA aptamers were covalently conjugated with a strand of siRNA, and the other siRNA strand was subsequently annealed to the aptamer-conjugated strand. These aptamer–siRNA conjugates were able to access HIV-infected cells and silence viral replication in vitro [58].

Viral vectors: natural siRNA carriers

Various recombinant viral vectors have been shown to be efficient in obtaining gene silencing for an extended period in a wide range of mammalian cells [59]. For example, an adenoviral vector encoding siRNA against pituitary tumor transforming gene 1 significantly inhibited the growth of the pituitary tumor transforming gene 1-overexpressing hepatocellular carcinoma cells in vitro and in vivo [60]. It was also demonstrated that the herpes simplex virus type 1-based amplicon vectors suppressed in vivo tumorigenicity of human polyomavirus BK-transformed cells (pRPc cells) [61]. Recombinant lentiviral vectors have also been frequently used to achieve in vivo gene silencing. In particular, lentiviral vectors containing the U6 promoter were found to be efficient in green fluorescent protein (GFP) silencing in vitro, resulting in ∼ 80% gene silencing at an average of one integrated vector genome per target cell genome. In addition, the U6 promoter was shown to be superior to the H1 promoter in achieving in vivo gene silencing and led to persistent GFP knockdown in the mouse brain for at least 9 months [62]. This indicates that lentivirus-mediated RNAi is a promising strategy for long-term gene silencing in vitro and in vivo. Other viral siRNA carriers such as retroviral vectors have not been intensively explored for their use in vivo [63–65]. Although viral vectors provide excellent tissue-specific tropism and high RNAi efficiency, safety concerns (e.g. insertion mutagenesis and immunogenicity) and difficulties with large-scale manufacture may limit the use of viral vectors for siRNA delivery in clinical setting [66,67]. Therefore, synthetic counterparts (nonviral vectors) have been more and more intensively explored as safe and effective alternatives that are easy to be prepared and can deliver large payloads of siRNA.

Nonviral carriers: Trojan horses for efficient, biocompatible and versatile siRNA delivery in vivo

Delivery of siRNA in its unmodified form has several advantages over using a chemically modified form. Unmodified siRNA possesses untouched RNAi capability (maximized RNAi per siRNA molecule) and does not require potentially inefficient and time/labor-consuming modification processes (cost-effective preparation). However, its highly anionic nature and the macromolecular size of siRNA necessitates using efficient carriers to overcome multiscale barriers. Unlike viral vectors, which deliver siRNA in the form of a viral genome, nonviral carriers deliver native siRNA, generate low immunogenicity and offer high structural and functional tunability. An ideally designed nonviral siRNA carrier with its desirable structural and functional multicomponents is depicted in Fig. 2.

Figure 2.

 An ideally designed nonviral siRNA carrier for efficient and targeted RNAi in vivo.

Liposomes and lipoplexes

One of the most significant advances in RNAi in vivo is successful knockdown of ApoB in nonhuman primates by systemically delivered siRNA in stable nucleic acid–lipid particles [68]. The siRNA–lipid complexes showed significantly enhanced cellular internalization and endosomal escape of siRNA. ApoB siRNA-carrying stable nucleic acid–lipid particles greatly reduced ApoB expression and serum cholesterol levels in monkeys when a clinically acceptable single siRNA dose of 2.5 mg·kg−1 was injected i.v. [68]. Importantly, expression of ApoB was silenced for at least 11 days. With addressing the high toxicity of the currently available liposomes for siRNA delivery in vitro and in vivo [69,70], cationic cardiolipin analog-based liposomes carrying c-raf siRNA inhibited the growth of breast tumor xenografts in mice [71]. Cationic liposomes formulated with anisamide-conjugated poly(ethylene glycol) effectively penetrated the lung metastasis of melanoma tumors in mice and resulted in 70–80% gene silencing after a single i.v. injection [72].

Further noticeable progress in siRNA delivery using liposomes is the use of neutral lipids for systemic siRNA delivery in order to address the toxicity of cationic lipids. For example, cyclin D1 (CyD1) siRNA was efficiently encapsulated in neutral phospholipid-based liposomes coated with hyaluronan [73]. The resulting siRNA-carrying liposomes were stable during circulation in vivo after i.v injection and suppressed leukocyte proliferation and cytokine secretion by type 1 T-helper cells. Another neutral dioleoyl phosphatidylcholine-based delivery system, which targets EphA2 [74] and focal adhesion kinase [75], demonstrated significantly inhibited tumor growth in an orthotropic ovarian cancer model in mice. The same type of liposome has also been reported to efficiently silence neuropilin-2 expression and inhibit the growth of colorectal cancer xenografts in the mouse liver [76].

Polymers and peptides

Nucleic acids such as siRNA are easily complexed with synthetic cationic polymers e.g., polyethylenimine (PEI), biodegradable cationic polysaccharide (e.g. chitosan) and cationic polypeptides [e.g. atelocollagen, poly(l-lysine) and protamine], via attractive electrostatic interactions. For example, i.t. injection of siRNA– atelocollagen complexes silenced luciferase expression in germ cell tumor xenografted in mice and inhibited tumor growth [77]. In another study, vascular endothelial growth factor (VEGF) siRNA–atelocollagen complexes significantly suppressed tumor angiogenesis and growth in a prostate tumor model in mice [78]. Intravenous administration of chitosan–RhoA siRNA complexes resulted in effective gene silencing in subcutaneously implanted breast cancer cells in mice [79]. In addition, intranasally administered chitosan–siRNA complexes efficiently silenced GFP expression in bronchiole epithelial cells in GFP-transgenic mice [29]. Tumor necrosis factor expression in systemic macrophages was silenced in mice after i.p. administration of chitosan/siRNA complexes, thus downregulating systemic and local inflammation [80].

Polyethylenimine is one of the most popularly investigated synthetic cationic polymers for nucleic acid delivery in vitro and in vivo. Polyethylenimine is very potent in transfection with its uniquely high buffering capability at an endosomal pH (proton sponge effect) which releases nucleic acid payloads into the cytoplasm [39]. c-erbB2/neu (HER-2) siRNA was delivered to subcutaneous tumors via i.p. administration of siRNA/polyethylenimine complexes and resulted in a remarkable reduction of tumor growth [81]. Pain receptors for N-methyl-d-aspartate were effectively knocked-down by intrathecal delivery of polyethylenimine-conjugated siRNA in rats [82]. Inhibited viral propagation in the lungs was also observed after deacetylated linear polyethylenimine/siRNA complexes targeting influenza nucleoprotein was retro-orbitally administered [83]. In another study, polyethylenimine-conjugated siRNA against secreted growth factor pleiotrophin reduced tumor growth and cell proliferation with no toxicity or abnormal animal behaviors after intracerebral administration in an orthotopic glioblastoma mouse model [84]. Overall, polyethylenimine seems to be a promising nonviral carrier for siRNA delivery in vivo, if its high toxicity and limited biodegradability are appropriately addressed.

Polypeptides, such as poly(l-lysine) and protamine, have also commonly been used to deliver siRNA. A sixth generation of dendritic poly(l-lysine) was employed to systemically deliver siRNA to silence ApoB expression without hepatotoxicity in hepatocytes of apolipoprotein E-deficient mice [85]. Protamine, a natural arginine-rich cationic polypeptide, condenses negatively charged nucleic acids and has been used as an efficient gene-delivery carrier [86]. An in vivo study demonstrated that complexes of siRNA and low molecular mass protamine, which possess membrane-translocating potency, were accumulated in tumors via i.p. administration and successfully inhibited the expression of VEGF, thereby suppressing the growth of hepatocarcinoma tumors in mice [87]. In addition, no noticeable increase in inflammatory cytokines, including interferon-α and interleukin-12, in serum was observed when the low molecular mass protamine/siRNA complexes were administered, indicating negligible immunostimulatory effects.

One of the fundamental concerns in using synthetic polymers for siRNA delivery in vivo is dose-dependent toxicity upon systemic administration. For example, polyethylenimine and poly(l-lysine) have been shown to trigger necrosis and apoptosis in a variety of cell lines [88,89]. The toxicity can be ameliorated by conjugation with biocompatible, hydrophilic polymers such as poly(ethylene glycol) or by removing excess (i.e., uncomplexed) cationic polymers. In gneral, natural cationic polymers (e.g. chitosan and protamine), which are biocompatible, biodegradable and nontoxic, are more desirable in siRNA delivery in vivo than synthetic polymers.

Targeted siRNA delivery in vivo

In order to achieve RNAi in vivo via systemic delivery, it is crucial for siRNA to be efficiently located in desired tissues/cells. This requires three important processes: prolong circulation in the body, high accessibility to target tissues and specific binding to target cells. Targeted siRNA delivery maximizes the local concentration in the desired tissue (maximized and localized silencing effects) and prevents nonspecific siRNA distribution (minimized unwanted effects in non-target tissues). For example, recent studies have reported cancer-targeted siRNA delivery using nanoparticles that specifically bind to cancer-specific or cancer-associated antigens and receptors [90,91].

Folate-conjugated siRNA carriers

One of the most popular target molecules in cancer-specific gene and drug delivery is the folate receptor [92]. Folic acid (FA) is needed for rapid cell growth, and many cancer cells overexpress folate receptors to which FA and monoclonal antibodies specifically bind [93]. FA can be easily conjugated onto the surface of liposomal and polymeric siRNA carriers with or without a poly(ethylene glycol) spacer [92]. For example, FA-conjugated polyethylenimine showed enhanced gene silencing via receptor-mediated endocytosis [94]. Chimeric survivin siRNA incorporated with bacteriophage phi29-encoded RNA and when further conjugated with FA suppressed the growth of nasopharyngeal carcinoma in mice, whereas control FA-free siRNA–phi29-encoded RNA hybrid did not affect tumor development [95]. As described earlier, RNA aptamer-mediated targeted siRNA delivery by direct conjugation with siRNA or tethering onto carriers has been a frequently adapted strategy.

Arg–Gly–Asp peptide-conjugated siRNA carriers

Arg–Gly–Asp (RGD) peptide targets tumor vasculature expressing αvβ3 integrin. Poly(ethylene glycol)ylated poly(ethylenimine) conjugated with RGD peptides was developed to selectively deliver VEGF siRNA to tumors [96]. In this study, i.v. injected polyethylenimine-poly(ethylene glycol)-RGD/siRNA complexes inhibited tumor angiogenesis and the growth of integrin-expressing murine neuroblastoma tumors in mice [96]. Systemic delivery (i.v. injection) and local delivery of poly(ethylene glycol)-polyethylenimine-RGD complexing VEGF siRNA also showed a significant inhibitory effect on virus-induced angiogenesis as well as the development of herpetic stromal keratitis lesions [97].

Antibody-conjugated siRNA carriers

Many studies have suggested that antibodies are good targeting modalities for targeted siRNA delivery in vivo, when careful selection of target antigen is made. Ideal antigens should be exclusively expressed or substantially overexpressed on target cells. Examples of antigens that have been used for cancer-targeted drug and gene delivery include HER-2 [98] and epidermal growth factor receptor [99]. For example, HER-2 siRNA-carrying liposomes decorated with transferrin receptor-specific antibody fragments (i.e. nanoimmunoliposome) silenced the HER-2 gene in xenograft tumors in mice, significantly inhibiting tumor growth [100]. An antibody fragment against an HIV gp160 has also been used for targeted siRNA delivery in vivo. siRNA linked to a protamine–antibody fusion protein, called F105-P, showed inhibited HIV replication in infected primary T cells [101]. Moreover, i.t. or i.v. injection of F105-P/siRNA complexes into mice successfully targeted gp160-expressing B16 melanoma cells. A synthetic chimeric peptide, which consists of nonamer arginine residues (9R) added to the C-terminus of a rabies virus glycoprotein peptide (29 amino acids) (RVG-9R), was able to specifically deliver siRNA to acetylcholine receptor-expressing neuronal cells after i.v. administration [102]. In addition, treating mice with Japanese encephalitis virus siRNA complexed with RVG-9R showed robust protection of the animals from lethal infection.

Intracellular siRNA delivery

In many aspects, siRNA delivery is similar to that of delivering other types of nucleic acids such as plasmid DNA, because they share most extracellular and intracellular barriers. However, several unique challenges in siRNA delivery make achieving efficient RNAi difficult compared with plasmid DNA delivery. First, the final target destination of siRNA is the cytoplasm, whereas plasmid DNA must be transported into the nucleus. This implies that siRNA should be rapidly released from its carrier upon endosomal escape. Second, overall RNAi efficiency is proportional to the number of siRNAs complexed with RNA-inducing silencing protein complex, whereas a successfully delivered single copy of plasmid DNA might be sufficient to express new transgene proteins. In other words, the maximum possible number of siRNA needs to be delivered in the cytoplasm in order to achieve the desired biological effects. Third, siRNA acts only once, whereas plasmid DNA can be replicated or even can be incorporated into the host chromosome [103] (short vs. permanent effects).

Cell-penetrating peptide-mediated siRNA delivery

Cell-penetrating peptides (CPPs), short cationic polypeptides with a maximum of 30 amino acids, have been extensively used to obtain enhanced intracellular delivery of a wide range of macromolecules [104]. CPPs have been shown to bind the anionic cell surface through electrostatic interactions and rapidly induce cellular internalization through relatively unclear mechanisms, although recent evidence shows that CPP-mediated internalization might be an endocytosis-mediated process [105,106]. Various CPPs, including TAT and MPG proteins from HIV-1 [107–110], as well as penetratin and polyarginine [111,112], have been employed for intracellular delivery of various proteins and nucleic acids.

Oligoarginine (e.g. 9 arginine, 9R), the simplest CPP, conjugated with cholesterol was shown to efficiently deliver siRNA to a transplanted tumor in mice [113]. It was also reported that HER-2 siRNA complexed with short arginine peptide was localized in perinuclear regions of the cytoplasm in vitro, further significantly inhibiting tumor growth of ovarian cancer xenografts [114]. Polyamidoamine dendrimer-TAT conjugated with bacterial magnetic nanoparticles was also used to deliver epidermal growth factor receptor siRNA to human glioblastoma cells in vitro as well as xenografts [115]. Another type of CPP, MPG-8, was also used to complex cyclin B1 siRNA, and the resulting complexes were further decorated with cholesterol for i.v. injection to the mice bearing human prostate carcinoma and human lung cancer xenografts [116]. The results showed efficient siRNA delivery in vivo at a low effective dose (0.5 mg·kg−1), indicated by inhibited tumor growth.

CPP-mediated cellular internalization via endocytosis requires additional molecules for facilitated cytosolic release of siRNA. For example, it was found that TAT–siRNA conjugates resulted in no gene silencing because they were entrapped in the endosomes even after efficiently entering cells [117]. Photostimulating fluorescently labeled TAT efficiently released TAT–siRNA conjugates from the endosome, resulting in enhanced gene silencing efficiency. Chloroquine and influenza virus-derived hemagglutinin peptide have also been frequently used to destabilize the endosomal membrane and enhance the cytosolic release of CPP-conjugated macromolecules [118–120].

Fusogenic or pH-responsive intracellular delivery of siRNA

Fusogenic peptides and lipids and pH-responsive lipoplexes and polyplexes have been used to ensure facilitated siRNA into the cytoplasm from the endosomes. For example, the incorporation of polypeptides derived from the endodomain of the HIV-1 envelope (HGP) or influenza virus fusogenic peptide (diINF-7) significantly promoted the liposomal fusion with the endosomal membrane, enhancing siRNA escape into the cytoplasm [40,121]. Similarly, equipping lipoplexes with fusogenic lipids, such as dioleoyl phosphatidylethanolamine (DOPE), was shown to facilitate the endosomal release of siRNA payload [122,123].

Stimuli-triggered macromolecule release from the mildly acidic endosome (e.g. pH 5.0–6.0) has been popularly investigated using a number of novel acid-responsive polymers [124–126]. For example, poly(ethylene glycol) shielding the surface of a highly fusogenic phosphatidylethanolamine lipid vesicles was cleaved upon acid hydrolysis of the vinyl ether bond, triggering fusion with the endosomal membrane [127]. A matrix metalloproteinase-cleavable and pH-sensitive GALA peptide was also used to link poly(ethylene glycol) and dioleoyl phosphatidylethanolamine (DOPE) lipid to obtain enhanced siRNA delivery specifically into cancer cells [128]. Highly efficient siRNA-mediated knockdown of luciferase expression was achieved in human fibrosarcoma cells in vitro and xenografted tumors using this method. Acid-degradable ketalized linear polyethylenimine significantly increased gene silencing efficiency via efficient cytosolic release with high resistance to serum and low cytotoxicity [129]. It was demonstrated that ketalized linear polyethylenimine/siRNA polyplexes were efficiently released into the cytoplasm upon acid-hydrolysis of ketal branches in the endosomes, followed by enhanced siRNA disassembly from ketalized linear polyethylenimine in the cytoplasm [129].

Conclusion

RNAi is an emerging therapeutic strategy and has been widely investigated. Despite a few promising clinical trials, effectively delivering siRNA in vivo remains a pivotal challenge in translating RNAi in the clinic as a conventional treatment option. A number of delivery systems and strategies have been developed to overcome multiscale extracellular and intracellular barriers to siRNA delivery in vivo, as summarized in Table 1. Chemically modified siRNA is stable against enzymatic degradation but can be cleared easily, generating potentially hazardous metabolites. Viral siRNA delivery raises several safety and preparation concerns such as immune responses and limited large-scale production. Nonviral siRNA carriers are efficient, safe and versatile in tackling the barriers in siRNA circulation, permeation into desired tissues, specific binding to target cells and optimized intracellular trafficking. Recent advances clearly indicate that interdisciplinary approaches using biology, chemistry and engineering play crucial roles in achieving efficient and targeted siRNA delivery in vivo.

Table 1.   siRNA delivery systems for RNAi in vivo. BCL-2, B-cell lymphoma 2; Cyb1, cyclin B1; CyD1, cyclin D1; DOPC, 1,2-dioleoyl-sn-glycero-3-phosphatidylcholine; DOPE, dioleoyl phosphatidylethanolamine; DOTAP, (N-[1-(2,3-dioleoyloxy)]-N-N-N-trimethyl ammonium propane); DPPE, dipalmitoyl phosphatidylethanolamine; DSPE, distearoyl phosphatidylethanolamine; FAK, focal adhesion kinase; HST-1/FGF-4, fibroblast growth factor; i.c.v., intracerebroventricular; i.p., intraperitoneal; i.v., intravenous; MMP-2, matrix metalloproteinase-2; NMDA, N-methyl-d-aspartate; NR2B, NMDA-R2B receptor subunit protein receptors; PAMAM, polyamidoamine dendrimer; PLK-1, polo-like kinase 1; PTTG1, pituitary tumor transforming gene 1; RVG, rabies virus glycoprotein; SNALP, stable nucleic acid-lipid particles; TNF-α, tumor necrosis factor-α; VEGF, vascular endothelial growth factor.
Delivery systemTarget geneIn vivo modelaDelivery routeRef.
  1. All the listed in vivo models involved a mouse model except Zimmermann et al. [68] and Tan et al. [82].

Cholesterol–siRNAApoBApoB transgenic micei.v.22
RNA aptamer–siRNAPLK-1, BCL-2Prostate tumor xenografti.t.57
Adenoviral vectorPTTG1Hepatoma tumor xenografti.t.60
Lentiviral vectorGFPGFP transgenic brainStereotactic62
Stable nucleic acid lipid particles (SNALP)ApoBMonkeysi.v.68
Cardiolipin analog-based liposomec-RafBreast tumor xenografti.v.71
DSPE–poly(ethylene glycol)–DOTAP–cholesterol liposomeLuciferaseB16F10 melanoma tumorsi.v.72
Hyaluronan–DPPE liposomeCyD1Gut inflammationi.v.73
Neutral DOPC liposomeEphA2Ovarian canceri.v.74
Neutral DOPC liposomeFAKOvarian canceri.p.75
Neutral DOPC liposomeNeuropilin-2Colorectal tumor xenografti.p.76
AtelocollagenHST-1/FGF-4 LuciferaseGerm cell xenografti.t.77
AtelocollagenVEGFProstate tumors xenografti.t.78
ChitosanEGFPTransgenic EGFP miceIntranasal29
ChitosanRhoABreast tumors xenografti.v.79
ChitosanTNF-αMicei.p.80
PolyethylenimineHER-2Ovarian tumor xenografti.p.81
PolyethylenimineNR2BNociception in ratsIntrathecal82
PolyethylenimineInfluenza nucleoproteinInfluenza virus infected-lungRetro-orbital83
PolyethyleniminePleiotrophin (PTN)Glioblastoma xenograftIntracerebral, i.p.84
Poly(l-lysine)ApoBMicei.v.85
ProtamineVEGFHepatocarcinoma xenografti.p.87
RGD–poly(ethylene glycol)–poly(ethylenimine)VEGFNeuroblastoma xenografti.v.96
RGD–poly(ethylene glycol)–poly(ethylenimine)VEGFCorneal neovascularizationSubconjunctival, i.v.97
HER-2-liposomes with histidine–lysine peptideHER-2Pancreatic tumor xenografti.v.100
HIV antibody–protaminec-myc, MDM2, VEGFB16 melanoma cells expressingi.t.101
 HIV envelopi.v. 
ArginineRVGNeuronal cellsi.v.102
Oligoarginine (9R) conjugated-water-soluble lipopolymer (WSLP)VEGFColon adenocarcinoma xenografti.t.113
Oligoalginine (15R)HER-2Ovarian tumor xenografti.t.114
TAT-PAMAMEGF receptorGlioblastoma xenografti.t.115
Cholesterol-MPG-8CyB1Prostate tumor xenografti.t.116
Lung tumor xenografti.v. 
DOPE-Cationic LipidLuciferaseMouse braini.c.v.123
GALA peptide–poly(ethylene glycol)–MMP-2 cleavable peptide-DOPELuciferaseFibrosarcoma xenografti.t.128

Acknowledgement

This work was supported by NSF CAREER Award (DMR-0956091) and a Council on Research Computing and Libraries Research Grant (UC Irvine).

Ancillary