Disease Allele Specific Gene Silencing

Excepting the extremely rare cases of X-linked dominant FHL1 mutations in males (Quinzii et al. 2008), patients with dominant disorders possess one mutant and one normal copy of their specific myopathy-related gene. As the underlying pathogenic events in these disorders are dominant gene mutations, simply reducing mutant allele expression may be therapeutic. For example, myotilin (MYOT) mutations cause dominant, progressive muscle disease clinically classified as LGMD1A, myofibrillar myopathy (MFM), or spheroid body myopathy (SBM; Foroud et al. 2005; Garvey et al. 2006; Hauser et al. 2000, 2002; Schroder et al. 2003; Selcen and Engel 2004). To date, 12 distinct point mutations were associated with dominant myotilinopathies in numerous families (Foroud et al. 2005; Garvey et al. 2006; Hauser et al. 2000, 2002; Schroder et al. 2003; Selcen and Engel 2004). Myotilin is a Z-disc protein expressed predominantly in skeletal and cardiac muscle.

Mutations cause myofibrillar aggregation and muscle weakness that is recapitulated in transgenic mice expressing a dominant human mutation (T57I; Garvey et al. 2006). Importantly, MYOT null mice are normal; they show no obvious muscle pathology or weakness, Z-disk proteins are unaltered, and animals live a normal lifespan (Moza et al. 2007). These data suggest that there may be a compensatory mechanism to counteract MYOT deficiency in mice. Whether MYOT absence is well-tolerated in humans is unknown. However, because LGMD1A phenotypes are recapitulated in an available mouse model, and MYOT absence produces no overt defects in mice, it may be an ideal target to demonstrate proof-of-principle for RNAi therapy of dominant muscle disorders.

In contrast to myotilinopathies, which may be an exception, most dominant muscular dystrophies may require specific silencing of the dominant allele. As normal copies of disease genes likely encode essential proteins, normal allele haploinsuffi-ciency may contribute to myopathic phenotypes as well. Loss-of-function contributions to dominant disease can be predicted from knockout mouse models and by examining genetic case studies, in which different mutations in the same gene give rise to dominant and recessive myopathies. For example, nemaline myopathy (NM) can arise from autosomal dominant or recessive TPM3 mutations (Corbett et al. 2005; de Haan et al. 2002; Durling et al. 2002; Ilkovski et al. 2008; Kee and Hardeman 2008; Laing et al. 1995; Lehtokari et al. 2008; Penisson-Besnier et al. 2007; Tan et al. 1999). Dominant NM patients have one mutant and one normal TPM3 gene copy, while human carriers of recessive alleles and TPM3+/- mice are normal, and TPM3-/- animals die as embryos (Lehtokari et al. 2008; Rethinasamy et al. 1998). These observations support two conclusions: only one normal TPM3 allele is required to maintain normal muscle, and gain-of-function TPM3 mutations are most likely the sole pathogenic event in dominant NM forms. Therefore, an RNAi strategy that specifically suppresses mutant TPM3 while leaving the normal allele untouched may improve myopathy in NM patients. Likewise, disease allele-specific RNAi therapies may be important for Caveolin-3-related myopathies, as normal Cav-3 gene dosage impacts muscle disease severity (Carbone et al. 2000; Galbiati et al. 2001; Minetti et al. 1998; Traverso et al. 2008). Specifically, severe LGMD1C is caused by autosomal recessive homozygous or dominant negative Cav-3 mutations resulting in complete or 97% Cav-3 loss. In contrast, different mutations resulting in 84 or 50% Cav-3 reductions produced mild hyperCKemia without muscle weakness, or normal phenotypes, respectively (Carbone et al. 2000; Galbiati et al. 2001; Minetti et al. 1998; Traverso et al. 2008). In both NM and LGMD1C examples, it would be advantageous to restrict gene knockdown to the affected allele while leaving the normal allele unperturbed. As many dominant myopathies are caused by single point mutations in one allele, the question arises: can inhibitory RNAs be designed to distinguish two transcripts differing by 1 base-pair? In short, the answer is yes. As previously discussed, perfect sequence complementarity between an inhibitory RNA and target mRNA causes message degradation; imperfect base-pairing leads to translational inhibition. However, this rule is not absolute. Complementarity does not ensure inhibitory RNA efficacy; not all inhibitory RNAs containing perfect homology with a target mRNA actually cause gene silencing. Conversely, more mismatch does not

Fig. 7.3 Hypothetical example of mutant allele-specific Cav-3 targeting. Wild-type (WT) and mutant T78K Cav-3 sequences are shown. The C to A Cav-3 mutation is located centrally within the miRNA sequence. This theoretical T78K-specific inhibitory RNA also contains a secondary peripheral mutation, as discussed in the text

Fig. 7.3 Hypothetical example of mutant allele-specific Cav-3 targeting. Wild-type (WT) and mutant T78K Cav-3 sequences are shown. The C to A Cav-3 mutation is located centrally within the miRNA sequence. This theoretical T78K-specific inhibitory RNA also contains a secondary peripheral mutation, as discussed in the text necessarily reflect reduced potency; miRNAs can have several mismatches with a target mRNA and still cause gene silencing, but a single nucleotide difference may be sufficient to prevent silencing altogether (Kurosawa et al. 2005; Lewis et al. 2005; Miller et al. 2003, 2005; Rodriguez-Lebron and Paulson 2006; Schwarz et al. 2006). Thus, well-designed inhibitory RNAs can specifically silence disease genes by distinguishing between normal and mutant alleles differing by one nucleotide. Although each allele-discriminating miRNA must be uniquely designed and empirically validated, some general guidelines can be followed. Specifically, the discriminating nucleotide should be placed centrally within the inhibitory RNA duplex and if sufficient disease allele-specific silencing is not produced, optimal specificity can be achieved by including additional peripheral mismatches in the inhibitory RNA sequence (Fig. 7.3).

7.6 RNAi Therapy for the Most Common Dominant Muscular Dystrophies

DM1 and FSHD are among the top three most common muscular dystrophies and both are dominantly inherited. Therefore, DM1- and FSHD-targeted treatments would potentially have the broadest benefit for patients with dominant muscle disease, making them logical candidates for RNAi therapy development. However, both disorders have complicated and unique etiologies that make them challenging, though not impossible, targets for RNAi treatment.

7.6.1. Myotonic Dystrophy Type 1

DM1 is caused by CTG trinucleotide repeat expansion in the DMPK 3' UTR, which causes nuclear retention of this toxic mRNA (Cho and Tapscott 2007). Patients develop myotonia leading to skeletal muscle weakness, and cardiac conduction abnormalities that often cause death in patients. DMPK+/- and -/- knockout mice both show skeletal and cardiac muscle sodium channel gating abnormalities that recapitulate conduction defects in human DM1 (Berul et al. 1999; Lee et al. 2003a; Mounsey et al. 2000). Older (7-11 month) DMPK+/- mice also show mild, variable sarcomeric disorganization, myofiber regeneration, and decreased force production (Reddy et al. 1996). Together, these phenotypes support that DMPK haploinsufficiency may contribute to some DM1 pathologies, which could complicate RNAi therapy for several reasons: (1) wild-type and mutant DMPK alleles are identical except for the 3' UTR trinucleotide repeat expansion. Therefore, therapeutic RNAi would theoretically knock down normal and mutant DMPK equally. Reducing normal DMPK could dampen beneficial effects caused by silencing the expanded mutant allele. (2) To target the mutant allele specifically, disease-linked polymorphisms, located outside the CTG repeat area, would have to be identified and (3) if possible, complete mutant DMPK knockdown would yield half the normal DMPK amount, potentially resulting in haploinsufficiency-related DM1 phe-notypes. One strategy to circumvent these potential problems would involve knocking down endogenous mutant and wild-type DMPK and simultaneously delivering a normal DMPK cDNA engineered with base changes that prevent its regulation by the therapeutic miRNA. A final issue regarding the feasibility of DM1-targeted RNAi therapeutics relates to sub-cellular localization of RNAi processes and mutant DMPK. Mutant DMPK transcripts are nucleus-sequestered, but recent conventional thinking was that RISC-mediated gene silencing only occurred in the cytoplasm, raising doubts about whether RNAi therapy could work for DM1 (Lee et al. 2002). However, several recent studies demonstrated that nuclear RISC exists, and that RNAi can reduce nuclear-localized transcripts, including 7SK and, importantly, DMPK (Langlois et al. 2005; Ohrt et al. 2008; Robb et al. 2005; Weinmann et al. 2009). Therefore, RNAi therapy for DM1 is feasible, but some complicating factors, discussed above, may have to be addressed to make it a therapeutic option in humans.

0 0

Post a comment