RNAi Therapeutics

RNAi molecules can be engineered to suppress any gene. Numerous strategies to design inhibitory RNAs have been developed and all share two common features: artificial RNAi molecules are double-stranded and comprised of sequences cognate to an mRNA of interest. Artificial inhibitory RNAs can be designed to mimic mature, pre-, or pri-miRNAs and will thus, upon delivery to cells, enter the miRNA pathway at different points (Fig. 7.1). There are three main classes of inhibitory RNAs (Fig. 7.2). (1) Small inhibitory RNAs (siRNAs) are in vitro synthesized, dsRNAs that are structurally identical to miRNA duplexes (Elbashir et al. 2001a). When delivered to cells, all siRNAs bypass the transcription and nuclear processing steps of the miRNA pathway. Some designed siRNAs are processed by Dicer (Rose et al. 2005), while others avoid this step and are immediately available to complex with RISC proteins after delivery to the cytosol. (2) Short hairpin RNAs (shRNAs) are structurally similar to stem-loop pre-miRNAs. They are typically designed to contain ~21 nt of paired stem sequence connected by an unpaired loop that is often derived from natural miRNA sequences (Paddison et al. 2002). ShRNAs are produced intracellularly, arising as transcripts from DNA expression cassettes using RNA pol III, and very rarely, pol II promoters. ShRNAs mimic Drosha-processed miRNAs and thus, following transcription, are immediately shuttled by Exp-5 to the cytoplasm for Dicer processing and incorporation into RISC. (3) Artificial miRNA shuttles resemble pri-miRNAs (Boudreau et al. 2009; Zeng et al. 2005). Like shRNAs, miRNA shuttles are transcribed from DNA expression cassettes, but are amenable to regulation by both pol II and pol III promoters. In this design, miRNA sequences required to direct Drosha and Dicer processing are maintained, but the natural, mature, 21-25 nt miRNA sequence is replaced by an inhibitory RNA sequence targeting the gene of interest. Thus, a natural miRNA is used to deliver an artificial siRNA. MiRNA shuttle transcripts are produced intracellularly and utilize all processing steps required for natural miRNA biogenesis.

Each of the three systems described above is capable of eliciting strong RNAi responses in vitro and in vivo. The key difference between siRNAs and shRNA/ miRNAs is duration of expression. In vitro synthesized siRNAs are transient and long-term disease-gene suppression requires repeated administration; expressed shRNAs or miRNA shuttles are longer lasting, and if delivered via an appropriate viral vector, may produce permanent gene silencing effects. Importantly, muscle-directed gene delivery systems are well-developed, especially those using adeno-associated viral (AAV) vectors, which have been used extensively in the last few years to deliver shRNA/miRNA to numerous tissues (Fechner et al. 2008; Grimm et al. 2006; Harper et al. 2005; Xia et al. 2004).

As described above, shRNAs and miRNAs differ in the level of processing required by endogenous miRNA biogenesis machinery. This differential processing has direct implications for how each is expressed. Because shRNAs are not Drosha processed, their 5' end must be defined by the start of transcription. This is important because Dicer binds the "Drosha-cut" end of the pre-miRNA and makes a defined cut ~21 nt downstream, which ultimately determines the sequence of the mature guide strand molecule (Fig. 7.2). As a result, shRNAs must be positioned near a promoter's transcription start site to ensure proper processing and gene silencing function. This restriction is not necessary for miRNA shuttles because Drosha processing, not transcription, defines the critical 5' Dicer binding site. As a result, artificial miRNAs can be expressed from any promoter. Moreover, several bifunctional expression vectors have been described, in which a coding gene and intron- or UTR-embedded miRNA

arise from the same pol II promoter-driven transcript (Du et al. 2006; Harper et al. 2006). Another difference between shRNAs and miRNAs is potential for nonspecific toxicity; miRNAs may be safer than shRNAs in vivo (Boudreau et al. 2009). ShRNAs were the first generation of plasmid- or vector-expressed artificial inhibitory RNAs used in vivo. Several studies have demonstrated shRNA efficacy for silencing disease genes and improving associated pathologies in, for example, models of neurodegen-erative disease and viral infection (Grimm et al. 2006; Harper et al. 2005; Li et al. 2003; Xia et al. 2004). However, a few recent studies have raised concerns about shRNA safety. Specifically, uncontrolled, high-level shRNA expression from constitu-tively-active pol III promoters caused liver failure and brain striatal loss in mouse models of hepatitis and Huntington's disease (HD), respectively (Grimm et al. 2006; McBride et al. 2008). This observed toxicity seems to be related to shRNA-induced saturation of endogenous miRNA biogenesis pathways, especially at the level of nuclear export, thereby interfering with natural miRNA function (Grimm et al. 2006). Importantly, lowering the dose of vector-expressed shRNAs in the liver, or using a less-powerful miRNA shuttle system in the brain, mitigated these toxic effects (Grimm et al. 2006; McBride et al. 2008). Both strategies ultimately led to significant gene silencing without overexpression associated toxicity. Although not all shRNAs are overtly toxic, and sufficient safety data regarding long-term artificial miRNA is lacking, miRNA shuttles may be safer than shRNAs simply because they are more efficiently processed and amenable to expression by tissue-specific, regulated, or weaker RNA pol II promoters, while shRNAs are dependent upon strong, constitutively-active pol III promoter expression. Regardless of the system used, RNAi therapy has shown promise in preclinical models of neurodegenerative disease, viral infection, and cancer, supporting its potential for treating dominant muscular dystrophies and other myopathies (Fechner et al. 2008; Grimm et al. 2006; Harper et al. 2005; Li et al. 2003; Mook et al. 2009; Saydam et al. 2005; Xia et al. 2004). These studies support its potential for treating dominant muscular dystrophies and other myopathies.

0 0

Post a comment