| |
Opportunities for treating chronic hepatitis B and C virus infection using RNA interference
|
| |
| |
Journal of Viral Hepatitis
Volume 14 Issue 7 Page 447-459, July 2007
P. Arbuthnot, V. Longshaw, T. Naidoo and M. S. Weinberg
Hepatitis B Virus Research Unit, Department of Molecular Medicine and Haematology, University of the Witwatersrand Medical School, Johannesburg, South Africa
RNAi, RNA interference; HBV, hepatitis B virus; HCV, hepatitis C virus; HCC, hepatocellular carcinoma; HBeAg, HBV e antigen; ORF, open reading frame; dsRNA, double-stranded RNA; siRNAs, short interfering RNAs; miRNAs, microRNAs; shRNA, short hairpin RNA; pdsDNA, partly double-stranded DNA; cccDNA, covalently closed circular DNA; MHI, murine hydrodynamic tail vein injection; AAV, adeno-associated virus; SNALP, stable nucleic acid lipid particle; IRES, internal ribosomal entry site; lhRNAs, long hairpin RNAs; PKR, dsRNA-dependent protein kinase receptor; TLRs, toll-like receptors; RIG-I, retinoic-acid-inducible gene-I.
Summary. Activating the RNA interference (RNAi) pathway to achieve silencing of specific genes is one of the most exciting new developments of molecular biology. A particularly interesting use of this technology is inhibition of defined viral gene expression. In this review, we discuss the potential application of RNAi to treatment of chronic hepatitis B virus (HBV) and hepatitis C virus (HCV) infection. Globally, these hepatotropic viruses are the most important causes of cirrhosis and liver cancer. Available treatments have their limitations, which makes development of novel effective RNAi-based therapies for HBV and HCV especially significant. Several investigations carried out in vitro and in vivo are summarized, which demonstrate proof of principle that HBV and HCV can be inhibited by RNAi activators. Challenges facing further development of this technology to a stage of clinical application are discussed.
Conclusions
During the past three years, several studies have demonstrated impressive knock down of HBV and HCV gene expression using RNAi. Although advances have been exciting and the technology is developing rapidly, some hurdles remain before the goal of therapeutic application of RNAi-based treatment of HBV and HCV infection is realized. Particularly important are the limitation of unintended effects, which include off-target silencing of cellular RNA, disruption of the endogenous miRNA pathway and unwanted immunostimulatory effects. Efficient delivery of RNAi-inducing nucleic acids, which has been an impediment to therapeutic nucleic acid transfer, is steadily improving. In addition, optimization of expression cassettes to include tightly regulated liver-specific promoters is the focus of several groups and will go some way to achieving precise dosing of expressed sequences. A particular difficulty with development of HCV treatment has been the lack of a suitable small animal model of virus replication. However, use of ingenious approaches such as HCV infection of SCID mice that have been grafted with human hepatocytes, will no doubt be an important precursor to clinical testing. With the concerted efforts from several quarters, considerable impetus has been given to the objective of developing treatment of HBV and HCV infection using RNAi and the next few years are likely to see considerable progress in the field.
Introduction
Globally, it is estimated that there are 387 million carriers of hepatitis B virus (HBV) [1,2] and 170 million people persistently infected with hepatitis C virus (HCV) [3]. Chronic infection with these viruses places individuals at very high risk for the serious complications of cirrhosis and hepatocellular carcinoma (HCC). Most HBV carriers are from sub-Saharan Africa, East and Southeast Asia as well as the western Pacific Islands. Between 25% and 40% of individuals who are chronic HBV carriers will develop HCC and/or cirrhosis [4]. The virus is transmitted parenterally, which is often through sexual contact in adults, perinatally in neonates or through poorly defined mechanisms of horizontal spread among children. Individuals who are infected at a very young age have a high (80-90%) chance of becoming chronic carriers. Conversely, transmission of HBV among adults is typically acute and self-limiting. Notably, it is the early onset carriers who are predisposed to HCC [5].
Eight HBV genotypes (A-H) have been described, which have distinct features of clinical course and geographical distribution. For example, HBV subgenotype A1, which is hyperendemic to South Africa, is associated with a particularly high risk for liver cancer that also occurs at a younger age [6]. Several mechanisms of HBV-induced HCC have been proposed and these include roles for integration of HBV DNA into the genome of hepatocytes, the potentially oncogenic HBx protein and the necroinflammatory hepatic disease that often accompanies infection (reviewed in [1]). Licensed treatments for HBV infection, which include interferon alpha (IFN-α), nucleoside (lamivudine) and nucleotide (adefovir) analogues, are partially effective [7,8]. A treatment response in HBV e antigen (HBeAg)-negative patients, such as is typical of sub-Saharan African carriers, is usually not durable. Prolonged treatment is required and is often complicated by emergence of resistance mutations in the tyrosine-methionine-aspartate-aspartate (YMDD) locus of the HBV polymerase open reading frame (ORF) [9,10]. Although other analogues such as adefovir may effectively suppress the YMDD mutants, long-term or lifelong therapy may be required to effect sustained suppression of HBV replication.
The six major genotypes of HCV, which differ in their sequence homology by more than 30%, have variable clinical course and geographical distribution (reviewed in [11,12]). Genotypes 1a and 1b are the commonest in Western Europe and North America, and genotypes 2 and 3 occur with lower prevalence in these regions. Genotype 4 is found commonly in Egypt, genotype 5 in South Africa and genotype 6 in Southeast Asia. The virus is typically transmitted percutaneously and perinatal transmission is uncommon. HCV infection is particularly common among human immunodeficiency virus type 1 (HIV-1)-infected individuals, intravenous drug users, their sexual partners and inmates of correctional facilities. Acute HCV infections are typically asymptomatic or associated with mild symptoms. However, the virus has a tendency to persist in 60-80% of cases, and may lead to cirrhosis and HCC. The mechanism of HCV-related hepatocarcinogenesis remains unclear and is likely to involve direct, indirect and genotype-related effects of the virus (reviewed in [13]). HCV core antigen has been implicated in causing liver cancer through induction of oxidative stress, mutations of growth regulatory genes and disruption of intracellular signalling to activate cell proliferation. Licensed therapies for HCV include pegylated (PEG) IFN-α and ribavirin, which are preferably administered in combination [14]. Response to therapy ranges from 45% to 80%. Improved success depends on several host and viral factors such as a low-serum viral load, less hepatic fibrosis, and infection with HCV genotypes 2 and 3.
Life threatening complications of infection with HBV and HCV are likely to remain significant global medical problems for several years. Although HBV vaccination has become widespread, it is expected that the number of HBV chronic carriers will remain high for many years to come. There is currently no vaccine available to prevent HCV infection. Recent discovery of the naturally occurring RNA interference (RNAi) pathway has led to considerable enthusiasm for development of novel effective nucleic-acid-based HBV and HCV therapies. The RNAi process involves specific and powerful silencing of gene expression through predictable complementary interaction between RNAi effectors and their targets. Earlier approaches to gene silencing using antisense and ribozyme technology have generally been found to be less effective than using RNAi-based approaches. Some recent publications have reviewed investigations that show potentially therapeutic use of RNAi for inhibiting HBV and HCV replications [15-17]. In this article, we summarize principal studies and we focus on considerations for development of RNAi-based technology to a stage of clinical application against HBV and HCV.
RNA Interference
RNAi is a conserved pathway that occurs in a diverse range of organisms, which includes plants, worms, flies and humans [18-20]. The process is initiated by double-stranded RNA (dsRNA) with sequence-specific homology to a target gene. Processing by the RNase III enzyme Dicer results in formation of short interfering RNAs (siRNAs) comprising 21-23 nucleotide sequences with 2 nucleotide overhangs at the 3' ends. One of the siRNA strands is incorporated into the cytoplasmic RNA-induced silencing complex (RISC) and functions as a guide for RNA degradation. Naturally, RNAi plays an important role in regulation of gene expression through the processing of endogenous microRNAs (miRNAs) (reviewed in [21-23]). These sequences are expressed as hairpin-like precursors from polymerase (Pol) II promoters and may be mono- or polycistronic. Effector miRNA sequences regulate expression of genes that are involved in important cellular processes, such as proliferation, differentiation and apoptosis. Typically miRNAs are not perfectly matched to their cognate targets and exert silencing through suppression of translation without cleavage of target mRNA. In a role that appears to be more important in lower eukaryotes, RNAi also protects cells from the deleterious effects of transposable gene sequences and viral gene expression. A recent interesting observation has been that miRNA miR-122, which is expressed specifically in the liver, facilitates replication of HCV [24]. MiR-122 functions atypically, as to date miRNAs have only been shown to induce post-transcriptional gene silencing. Nevertheless, this effect is mediated through interaction of miR-122 with the viral 5'-non-translated region (5'-NTR) and apparently does not modulate viral translation or RNA stability.
Exogenous silencing sequences, which are designed to mediate degradation of homologous target sequences, are typically either synthetic siRNA duplexes or RNA derived from intracellular promoter-based transcription. Intracellular generation of interfering RNA molecules is usually achieved by inserting a short hairpin RNA (shRNA)-encoding DNA sequence downstream of a Pol III promoter [25-28]. shRNA sequences enter the RNAi pathway as Dicer substrates, and mimic the processing of endogenous miRNAs. The small nuclear RNA U6 or human H1 promoters are the most commonly used. Unlike Pol II promoters, these transcription regulators have the advantage of containing all of their cis sequence elements, with the exception of the first transcribed nucleotide, upstream of the transcription initiation site. The Pol III termination is typically defined by a poly dT sequence (usually dTdTdTdT) in the sense sequence of the transcription template. When compared with synthetic sequences, DNA cassettes that encode interfering RNA have the benefits of obviating problems of RNase contamination and high cost of synthesis. However, synthetic RNA has the advantage of easier cellular delivery as well as the ability to regulate the dosage of silencing sequences precisely.
Several algorithms have been described to assist with optimizing RNAi effectors, and general properties of effective siRNA/shRNA sequences have been described, which help with design [29-31]. For example, to influence bias for strand selection, it is preferable that the intended guide sequence is less tightly bound (i.e. A:U rich) to its complementary strand at its 5' end within a siRNA duplex [32]. Moreover, target sites within RNAs that form strong secondary structures are known to be refractory to siRNA-mediated silencing [33,34]. Nevertheless, ensuring specificity of action and exclusion of unwanted effects usually requires additional empirical assessment. Unintended immunostimulation, cross-reaction with cellular RNA and disruption of the endogenous miRNA pathway may complicate development of RNAi-based therapy. These topics are discussed in the following sections in more detail.
Hepatitis B Virus Genome And Replication
The HBV, which is the prototype member of the Hepadnaviridae family of hepatotropic mammalian and avian viruses, is a small, noncytopathic, enveloped partly double-stranded DNA (pdsDNA) virus with a genome size of 3.2 kb [35-37]. The viral genome is remarkably compact and encodes four primary ORFs: core (C), polymerase (P), envelope (surface, S) and X (HBx) (Fig. 1). Following viral entry into the hepatocyte, pdsDNA is released from the viral nucleocapsid and is converted to covalently closed circular DNA (cccDNA). Within the hepatocyte nucleus, cccDNA serves as a template for transcription of capped and polyadenylated viral mRNA. Naturally, establishment of HBV infection occurs after a lag phase of 4-6 weeks and reaches a peak of 108-1013 viral particles/mL of serum. HBV spreads slowly in vivo, but is capable of infecting all hepatocytes [38,39]. Activation of the adaptive and innate immune responses is important for clearance of the virus [40]. HBV counters the adaptive immune response through induction of immune tolerance and generation of antibody decoys. Evasion and inhibition of the innate immune response are also likely to occur. Transcription is unidirectional and is initiated from individual promoters (preC/pregenomic, S1, S2 and X) to result in formation of preC/pregenomic, preS1, preS2/S and X mRNA (Fig. 1). The single polyadenylation signal terminates transcription of all the HBV transcripts at a common 3' end. Two enhancers (enhancer I and enhancer II), as well as cis-acting negative regulatory elements, also play important roles in the regulation of viral gene transcription [41]. The 3.5 kb preC/pregenomic mRNAs serve as the template for the precore/core, core and polymerase proteins in addition to being the reverse transcription substrate [42]. The precore/core protein is processed in the endoplasmic reticulum and secreted as the HBeAg. This antigen may function as an inducer of immunotolerance to inhibit the adaptive immune response. When detected in the serum of infected individuals, HBeAg is associated with a high viral replication rate and improved response to therapy with nucleoside/nucleotide analogues. The 2.4 and 2.1 kb mRNAs encode the surface proteins, and the small 0.9 kb transcript is the template for translation of HBx. Several lines of evidence implicate HBx in hepatocarcinogenesis and this viral protein may also inhibit the host innate immune response by suppression of proteasome-dependent degradation of viral particles [43,44]. The remarkably compact arrangement of the HBV genome with its limited sequence flexibility, together with the requirement for reverse transcription of pregenomic RNA during replication, make the virus well suited to developing RNAi-based therapy that involves nucleic acid hybridization to a viral target.
Use of RNA Interference Against Hepatitis B Virus Infection
Several sites of the HBV genome have been targeted with synthetic and expressed shRNA. To test the antiviral efficacy of RNAi sequences, use of replication competent plasmids that encode greater than genome length viral sequences has been made. These vectors can be readily introduced into cells in culture or hepatocytes in vivo using standard procedures of transfection or murine hydrodynamic tail vein injection (MHI). As there is no convenient small animal model of HBV infection, transgenic mice have also been used to simulate virus replication that occurs in HBV chronic carriers. Although HBV replication in these animals is robust, infection of hepatocytes does not occur, because murine cells lack the as yet unidentified HBV receptor.
In an early study using RNAi against HBV, McCaffrey et al. [45] assessed the anti-HBV efficacy of a panel of six U6 promoter cassettes that encoded shRNAs. The most effective sequence, which targeted a region of the surface and overlapping polymerase ORF, inhibited HBV surface antigen secretion by 94% in transfected cultured cells, and 85%in vivo in the MHI model. Inhibitory effects were observed in normal (C57BL/6) as well as immunocompromised NOD SCID mice, which indicate that the expressed shRNAs have a direct effect that is not dependent on an antigen-dependent immune response. Since this early report, others have demonstrated varying degrees of inhibition of HBV replication by expressed shRNA sequences [46-52] and one study showed antiviral synergy between lamivudine and shRNA sequences in a cell culture model [49]. Although these studies show promise for use of expressed sequences against HBV, emergence of RNAi escape mutants is an important consideration when developing this approach to therapy [53].
Three studies have reported efficiency of expressed shRNA in HBV transgenic mouse models [50,51,54]. In these mice, hepatotropic recombinant adenovirus vectors expressing shRNAs from Pol III promoters effected successful inhibition of viral replication [50,51]. Sustained suppression of markers of HBV replication was observed for 2-4 weeks after a single bolus dose of viral vector. However, the importance of optimizing dose of expressed sequences was highlighted in a recent study that showed fatality in HBV transgenic mice that were treated with shRNA-producing recombinant adeno-associated virus (AAV) type 8 vectors [54]. A high dose of recombinant vector expressing a 25-mer anti-HBV shRNA consistently caused death of treated mice, whereas low dose of a 19-mer shRNA vector efficiently suppressed markers of viral replication. Collectively, these studies provide important evidence that HBV is susceptible to gene silencing using RNAi and in particular when using a model of established ongoing replication that simulates the situation in HBV chronic carriers.
Highly effective knock down of HBV replication by synthetic siRNA has also been shown in various studies. A siRNA duplex that targeted sequence nucleotides 9-27 from the surface ORF initiation codon was found to be particularly effective against HBV without a requirement for HBV DNA synthesis [55]. This property is distinct from anti-HBV nucleoside or nucleoside analogues, which act on the viral DNA polymerase to have their therapeutic effect. Efficacy of surface ORF-targeted siRNAs was confirmed in other studies [56,57] and improved knock down by repeated siRNA transfection of cells in culture was also reported [58]. Another interesting observation has been that siRNAs are effective against a HBV target that includes the sequences encoding the YMDD polymerase gene mutation that is characteristic of lamivudine-resistant HBV strains [59]. Recently, Morrissey et al. [60] showed potent and persistent anti-HBV activity of chemically modified siRNAs that were administered intravenously within a stable nucleic acid lipid particle (SNALP) formulation. The chemical modifications included deoxynucleotide residues, incorporation of 2' fluoro C and U as well as 2' O methyl A and G. Improved efficiency of the complexes is likely to be a result of stability of these effector sequences in vivo and also the abrogation of nonspecific immunostimulatory effects (see below). No adverse hepatic and haematological toxic effects caused by these modified sequences were observed in mice.
In summary, comparison of anti-HBV activity of synthetic and Pol-III-derived shRNAs shows that both types of sequences effectively inhibit HBV gene expression. While synthetic siRNA duplexes may have more rapid and dose-controlled effects [55], the effects of expressed shRNAs may be more sustained. Differences between these two classes of interfering RNAs may be important for therapeutic application and this topic is discussed below in more detail.
Hepatitis C Virus Genome and replication
The HCV is a member of the Flaviviridae family of viruses and is classified within the Hepacivirus genus. The virion of HCV is enveloped and contains a single-stranded uncapped RNA genome of 9.6 kb with sense polarity (reviewed in [61]). An internal ribosomal entry site (IRES), responsible for initiating translation, is located within the 5'-nontranslated region (5'-NTR) of the HCV genome. One ORF is encoded by the HCV RNA, which is flanked by the 5'-NTR and 3'-NTR located upstream and downstream, respectively (Fig. 2). A large precursor polyprotein is processed by proteolytic cleavage to form 10 viral proteins, which are the core, E1, E2, p7, NS2, NS3, NS4A, NS4B, NS5A and NS5B HCV proteins. Host cellular and virally encoded proteinases cleave the polypeptide precursor to release individual viral proteins. Functions have been ascribed to most of the viral proteins. Envelope proteins are encoded by E1 and E2, the viral core is derived from the C sequence. Nonstructural proteins have roles in assembling the viral proteinase machinery (NS2, NS3 and NS4A), forming the specialized membrane compartment for viral replication (NS4B) and RNA-dependent RNA polymerase (NS5B). The hydrophobic p7 protein is thought to function as a viroporin, which is required for the late stages of viral assembly. In addition, a frameshift (F) protein has recently been described, which is generated by translation initiation at the core AUG and ribosomal frameshifting at approximately codon 11 [62-64]. This protein is expressed during natural infection, but its function is yet to be determined. Unlike with HBV, the entire HCV life cycle is cytoplasmic. HCV replication involves the formation of minus RNA and dsRNA intermediates within the membranous web. Although HCV dsRNA activates signal cascades of the innate immune response, cellular antiviral effects are countered by the E2, NS3 and NS5A proteins [65-68]. Sequence heterogeneity and rapid evolution of quasispecies are characteristic of HCV infection. This results from the high error rate of the RNA-dependent RNA polymerase encoded by NS5B. Sequence variability is not evenly spread throughout the genome, and the 5'-NTR sequence appears to be particularly well conserved [61]. Structural characteristics of this region, which determine its function as an IRES, probably impart sequence conservation. The HCV and HBV kinetics of replication in vivo contrast significantly [40]. HCV expansion occurs rapidly after infection and reaches a lower maximum viraemia of about 106 genome equivalents per millilitre serum in chimpanzees. In addition, sensitive analysis indicates that HCV replicates actively in only a few hepatocytes or at a low level throughout the liver [69].
Propagation of HCV in cell culture has been difficult to achieve, and successful reproduction of HCV replication in vitro has only recently been described [70-72]. Subgenomic replicon systems have been used successfully to study aspects of HCV replication and assess efficacy of antiviral therapeutic agents (reviewed in [73,74]). These replicons lack the structural sequences that are required to produce infectious virions, and have mutations that enable propagation in cell culture. Currently available models of HCV infection in vivo are limited and include the chimpanzee [40] and chimaeric immunodeficient mice that are grafted with human hepatocytes [75]. To date, most studies aimed at assessing RNAi knock down efficacy against HCV have used replicons or, alternatively, viral sequences that are fused to appropriate reporter genes.
Hepatitis C Virus As A Target For RNA Interference
As an RNA virus that replicates in the hepatocyte cytoplasm, HCV is a prime candidate for RNAi-based treatment. However, the high degree of heterogeneity of viral sequences has been a particularly significant obstacle to developing antiviral RNAi effectors. The 5'-NTR of HCV is more conserved than other regions of the HCV genome and has been the favoured HCV target of several studies [76-80].
As with HBV, HCV has been the target of both synthetic and expressed activators of the RNAi pathway and effective inhibition of markers of virus replication has been described in several studies. In one of the earliest investigations that employed RNAi against HCV, Yokota et al. [80] used both synthetic and expressed RNAi effectors against the 5'-NTR. Approximately 80% suppression was observed in a replicon model of the virus. In other studies targeting the 5'-NTR, naked shRNA-mediated inhibition of a HCV IRES reporter gene fusion cassette [79] and also dose-dependent specific reduction in HCV RNA with synthetic siRNA were demonstrated [81]. Endonuclease-prepared siRNAs (esiRNAs) and shRNAs expressed from pseudotyped retroviral vectors were used by Kronke et al. to target various HCV sequences [76]. Domain IV regions of the IRES were found to be a particularly good target for HCV gene silencing and this observation confirmed the findings of Yokota et al. [80]. Efficacy of shRNA sequences against the 5'-NTR was not observed in a study by Takigawa et al. [78]. This report demonstrated that in both plasmid and lentivirus expression systems, shRNAs against NS3-1 and NS5B suppressed the HCV replicon most efficiently, while slight suppression was observed for similar vectors that targeted the 5'-NTR. Again using both synthetic and expressed siRNAs, significant inhibition of virus gene expression was achieved in replicon models when targeting NS3 and NS5B sequences [82,83]. The antiviral effect of the siRNAs was independent of induction of the IFN response or any influence on cell cycle progression [83]. In a study using the MHI model in which a reporter gene was fused to a HCV target, McCaffrey demonstrated knock down with expressed shRNA [84]. Although the reports described earlier show positive results, a major concern remains on the ability of the virus to accumulate evading nucleotide sequence mutations. Not surprisingly, a recent report showed emergence of resistant HCV RNAs after several treatments of replicon-expressing cultured cells with siRNA targeted against the NS5B coding region [85]. The siRNA-resistant replicons showed point mutations within the siRNA target sequence. Resistant replicons were, however, sensitive to an siRNA that targets another part of the genome, suggesting that use of siRNA combinations limits evolution of escape mutants. Recent studies showed that modified long hairpin RNAs (lhRNAs) comprising 50-100 bp with multiple G:U pairings in the stem sequence are capable of silencing hepatitis C virus targets in cell culture [86,87]. This interesting observation implies that by targeting a greater viral sequence, viral escape would be minimized. Using the lhRNA approach, silencing of multiple genotypes may be possible and the probability of effective silencing by one of the several lhRNA-derived siRNAs is likely to be improved. Importantly, the IFN response was not induced by these lhRNA sequences and although the mechanism is not yet clear, it has been proposed that wobble base pairs in long duplex RNA may enable evasion of this off-target effect.
Challenges for developing RNA-interference-based hepatitis B and C virus therapy
The relative ease with which potent silencing of HBV and HCV sequences can be achieved using effectors of RNAi has led to considerable enthusiasm for developing this technology for therapeutic application. Although the studies summarized earlier augur well for this approach, some important challenges need to be addressed before clinical application of RNAi antivirals. These and other cautionary considerations are described below.
Immunostimulatory effects of short interfering RNAs and expressed hairpin sequences
Presence of dsRNA within a cell, such as may be produced during viral replication, may be sensed as unwanted gene activity and lead to release of inflammatory cytokines and activation of the IFN response. Although limited immunostimulation may aid an antiviral effect of siRNA/shRNA, uncontrolled induction would cause adverse effects and may potentially limit any therapeutic applicability of RNAi [88-92]. Two cellular mechanisms of induction of the IFN response by dsRNA have been described: activation of cytoplasmic pattern recognition receptors such as dsRNA-dependent protein kinase receptor (PKR) and interaction of RNA with toll-like receptors (TLRs) (Fig. 3). Stimulation of TLRs leads to a cascade of events such as activation of myeloid differentiation factor 88, which is followed by transcription activation by NFƒÈB, IRF3 and IRF7 and resultant increased expression of a range of genes that includes inflammatory cytokines (e.g. TNFα, IL6 and IL12) and IFNs (reviewed in [88,90-93]). Type 1 IFNs may also activate JAK-STAT pathways to cause autostimulatory positive feedback and amplify the response in other cells. IFN pathway activation may also lead to phosphorylation of the α subunit of translation initiation factor 2α (eIF-2α) and oligoadenylate-synthase-1-mediated activation of RNaseL (Fig. 3). Resultant inhibition of protein synthesis and degradation of cellular mRNA may lead to programmed cell death (apoptosis).
As the innate immune response has become better understood, several immunostimulatory properties of RNAi effectors have emerged. The type of effector sequence that is used to activate RNAi, i.e. expressed sequences or chemically synthesized siRNAs, has a bearing on the mechanism of inducing nonspecific effects [94]. Synthetic siRNAs that are longer than 30 bp [95], possess 5' triphosphates [96] and lack 2 nucleotide 3' overhangs [97] have been shown to be immunostimulatory. In addition, certain edangerf motifs (e.g. GU rich sequences, 5' GUCCUUCAA 3' and 5' UGUGU 3') may induce immunostimulation through activation of endosomal TLR3, TLR7 and TLR8 [98]. Localization of these TLRs within endosomes is likely to have evolved to enable discrimination of normal cellular RNA from exogenous potentially harmful sequences. Thus, cellular recognition of danger motifs depends to some extent on the presence of the RNA in the endosome as is usually the case during delivery of synthesized siRNA to cells with nonviral vectors. Recently, chemical modifications of siRNAs have been described, which attenuate immunostimulatory effects [99]. For example, incorporation of 2'O-methyl groups diminishes immunostimulation and is thought to result from reduced interaction with endosomal TLRs. This approach has been used successfully in vivo to counter HBV replication with siRNAs without induction or release of interleukins and inflammatory cytokines [60]. RNA that is derived from exogenous DNA is transcribed in the nucleus before export to the cytoplasm and therefore does not typically pass through the endosomal compartment to activate TLR3, TLR7 and TLR8. However, DNA expression cassettes themselves may mediate an immunostimulatory effect through the activation of TLR9 by unmethylated CpG islands (reviewed in [100]).
PKR and retinoic-acid-inducible gene-I (RIG-I) are the best characterized cytoplasmic receptors that activate an immunostimulatory response (Fig. 3) [88-92]. RIG-I binds duplex RNA in the cytosol and activates downstream effects after unwinding the double strand and signalling via its caspase recruiting domain [97]. Importantly RIG-I is capable of distinguishing endogenous Dicer products from foreign dsRNA, and immunostimulation is induced in the absence of 2 nt 3' overhangs. PKR is typically activated by binding to duplex RNA that is longer than 30 bp to activate signalling cascades. Both PKR and RIG-I induce downstream activation of IFN-β [88-92].
Nonspecific interaction of silencing molecules with cellular sequences
Cross-hybridization of HBV/HCV interfering RNA with transcripts that have partial sequence identity to the intended target may contribute to nonspecific effects. A less stringent requirement for exact matching of the full length of the siRNA sequence to its cognate sequence significantly diminishes the specificity of siRNA effects and will have important implications for developing RNAi-based therapy. Direct silencing of nontargeted sequences with as few as 11 contiguous complementary sequences has been reported [101]. Careful analysis of unintended silencing of cellular transcripts will need to be undertaken as a prerequisite for using RNAi-based therapy against HBV and HCV infection. It is likely that detailed microarray analysis to characterize acceptable off-target effects will become commonplace in development of therapeutic RNAi effectors.
Optimizing delivery vectors
One of the most important challenges for development of RNAi-based HBV/HCV therapy is optimization of delivery methods. To date, both viral and nonviral vectors have been used to deliver sequences to the liver in vivo. For safety and ease, systemic administration has generally been preferred to intrahepatic/intraportal injection of vectors. The most promising viral vectors have been found to be recombinant adenoviruses and AAVs [16]. Both adenoviruses and AAVs are capable of transducing liver cells in vivo with high efficiency. Adenoviruses have been used successfully in two studies to deliver sequences that silence HBV gene expression in transgenic models of HBV replication [50,51]. Using an indirect approach, adenoviral vectors expressing RNAi effectors against cellular HCV replication cofactors effectively silenced endogenous genes and inhibited markers of HCV replication in Huh-7 cells [102]. Importantly, toxicity of recombinant adenoviruses and an immune response to the viral antigens need to be addressed before clinical application. Adenoviruses induce strong innate and adaptive immune responses that may cause toxicity and limit repeated administration [103]. To reduce immune responses to adenoviruses, recent studies have used PEG and helper-dependent gutless vectors. Addition of PEG diminishes protein-protein interactions to inhibit immunostimulation. Importantly, these vectors remain capable of efficient gene delivery in vivo [104]. Helper-dependent adenoviral vectors have been modified to exclude all viral sequences except for the terminal repeat and packaging elements [105]. These vectors are more difficult to propagate than the first generation vectors but have an improved safety profile and are capable of effecting long-term (>2 years) gene transfer in vivo.
The AAVs have shown promise as vectors for hepatotropic delivery of RNAi effectors and there are a number of properties that make these vectors attractive. These include the lack of demonstrated pathogenicity associated with AAV infection and also the relative ease with which recombinant vectors can be propagated [106]. Double-stranded AAVs have been useful for improving the efficiency with which vectors transduce cells. More than 100 AAV serotypes, which have different properties, have been described. AAV-2 is the best studied and is regarded as the prototype vector. The AAV-8 vector is of particular interest for HBV and HCV therapy because of its hepatotropic properties. Both AAV-2 and AAV-8 have been used successfully to inhibit HBV replication [54,107].
To date, most studies in vivo, which have used nonviral vectors for delivery of anti-HBV/HCV sequences have incorporated synthetic siRNAs rather than DNA expression cassettes. The siRNAs are smaller than DNA expression cassettes, which makes their delivery across the cell membrane to the site of action in the cytoplasm easier to effect. DNA expression cassettes face an additional barrier to efficient delivery, which is the need to be transported across the nuclear membrane to the site of transcription. Nonviral vectors (SNALPs) have been used successfully to deliver anti-HBV synthetic siRNAs in an MHI model of HBV replication [60]. SNALPs were administered after resolution of hepatitis caused by MHI. Efficiency of siRNA delivery was apparently not affected by damage to hepatocytes, which was caused by the hydrodynamic injection procedure. To date, there have been no reports on the efficiency of siRNA-mediated inhibition in more stringent models of HBV replication such as HBV transgenic mice. However, use of SNALP vectors was shown to knock down endogenous hepatic gene expression in primates and suggests that this vector technology will have clinical applicability [108]. Recently developed nonviral vectors are capable of avoiding uptake by the reticuloendothelial system and may have the advantage over viral vectors of limited induction of an adaptive immune response.
Alternative targets
Traditionally, RNAi effector sequences have been targeted to HCV and HBV genes to effect silencing of viral replication. As summarized earlier, this technique has successfully achieved viral knock down. Alternative approaches aimed at silencing of host genes that are required for viral replication or involved in mediation of pathological consequences of viral infection also have been reported [102,109,110]. This approach has the important advantage of circumventing the problem of viral escape, which is particularly important with persistent HCV infection. Synthetic siRNA and adenovirus-delivered effectors specific for La autoantigen (La), polypyrimidine-tract-binding protein and human VAMP-associated protein of 33 kDa substantially blocked HCV replication in Huh-7 cells [102,109]. In another study aimed at facilitating the HCV silencing effects of cyclosporin A, RNAi sequences directed against the cyclophilins suppressed HCV replication significantly [110]. Cyclophilins are the intracellular ligands of cyclosporin A and specific blockade of cyclophilin expression represents an interesting anti-HCV approach. Endogenous miR-122, which has recently been shown to be required for HCV replication, is another cellular target that may be silenced to counter viral infection [24]. The Fas cell death receptor, known to mediate T-cell hepatocyte toxicity, caused by viral infection, was efficiently silenced using RNAi [111]. In this study, anti-Fas siRNAs also countered the severe hepatotoxicity that was induced by administration of Fas antibodies. Similarly, inhibition of caspase 8 which is a key enzyme in death-receptor-mediated apoptosis, prevented acute liver failure in mice [112]. Although promising, it is likely that countering inflammation caused by viral infection will require administration of additional therapeutic agents that inhibit viral replication directly. RNAi-based treatment of chronic HBV/HCV infection is likely to require administration over a long period, and whether silencing endogenous genes results in long-term unwanted consequences remains to be established.
Regulating dose
The recent demonstration that saturation of the endogenous miRNA processing machinery by exogenous expressed RNAi-inducing sequences causes serious toxicity in hepatocytes is an important caveat for the development of RNAi-based HBV and HCV treatment [54]. Grimm et al. [54] showed that expressed shRNA sequences were capable of usurping and disrupting essential endogenous miRNA-mediated cell functions. Their results also indicate that activity of exportin-5, a nuclear membrane karyopherin required for transport of precursor miRNA to the cytoplasm, is one of the rate-limiting steps of the miRNA pathway. Administration of synthetic siRNAs instead of expressed shRNA would bypass this step and may be a means of evading the harmful effects of saturating exportin-5 function. However, it remains to be determined whether cytoplasmic RNAi functions are similarly inhibited by downstream RNAi effector sequences. Although strong evidence is presented for the dose of shRNA-expressing vectors in causing toxicity, other variables may also contribute to the toxic effect. For example, the length of the stem sequence of the hairpin may play a role: shorter stem regions of 19 bp seemed less harmful than the longer 25 bp shRNA duplexes [54]. Improvement of efficacy and specificity of expressed shRNAs will involve the optimization of promoter sequences that express shRNAs. When using constitutively active Pol III promoters, such as U6 and H1, regulation of dosage and intracellular copy of shRNAs is difficult to achieve. Tissue-specific and inducible Pol II promoters may be a means of refining transcription control and therefore dosage of expressed RNAi effector sequences.
Mode of administration
Use of RNAi-based treatment as a prophylactic to prevent infection of hepatocytes has also been considered [16,85]. This is potentially useful for ex vivoevaccinationf of a donor liver against HBV/HCV. Employing RNAi to confer HBV/HCV immunity before liver transplant may afford protection against reinfection in the organ recipient. Although this approach is effective for protection of haematopoietic cells from HIV-1 infection [113,114], the feasibility of such a strategy is some way from being established for liver transplantation.
Targeting multiple viral sequences with RNA interference
An important concern of applying RNAi to the treatment of virus infections, particularly HCV, is the ability of some viruses to mutate and evade silencing effects of siRNA or shRNA. Selection of RNAi escape mutants has been reported in vitro for poliovirus [115], HIV-1 [116,117] and HCV [85]. However, HCV-resistant sequences were susceptible to silencing by siRNAs that targeted alternative sites [85]. Using vectors that deliver several siRNAs or expressed hairpin cassettes that target multiple viral sequences is potentially useful to overcome viral escape. A vector that produces a lhRNA Dicer substrate, which generates multiple siRNAs, has been shown to be effective against HCV [86,87]. Although use of dsRNA Dicer substrates with a duplex region greater than 30 bp may be complicated by the induction of the IFN response in mammalian cells [95], recent studies have showed that expressed hairpins have an attenuated effect on activation of the IFN response [94]. Therapeutic sequences that comprise or generate multiple siRNAs would have the advantage of limiting escape and targeting a range of sequences found in different viral genotypes or quasispecies. However, a concern of this approach is that unwanted effects such as the silencing of nontargeted genes and disruption of the endogenous miRNA pathway may be caused. In addition to using multiple RNAi effectors for targeting different sites within a viral genome to limit escape, combination with established licensed drugs that have a different mechanism of action may improve efficacy. Such synergy has been demonstrated when using anti-HBV sequences in conjunction with lamivudine [49]
Acknowledgements
The South African Innovation Fund, Griffin Trust and Poliomyelitis Research Foundation support research conducted in the authors' laboratory.
|
|
| |
| |
|
|
|
