Gene Therapy (2010) 17, 486–493 & 2010 Macmillan Publishers Limited All rights reserved 0969-7128/10 $32.00 www.nature.com/gt ORIGINAL ARTICLE AAV-mediated knockdown of Peripherin-2 in vivo using miRNA-based hairpins A Georgiadis1, M Tschernutter1, JWB Bainbridge1, SJ Robbie1, J McIntosh2, AC Nathwani2, AJ Smith1 and RR Ali1 1Department of Molecular Therapy, Institute of Ophthalmology, University College London, London, UK and 2Department of Haematology, University College London, London, UK Gene therapy for inherited retinal degeneration in which The results show that an miRNA-based shRNA can efficiently expression of a mutant allele has a gain-of-function effect on and specifically silence Prph2 in vivo as early as 3 weeks after photoreceptor cells is likely to depend on efficient silencing AAV2/8-mediated subretinal delivery, leading to a nearly 50% of the mutated allele. Peripherin-2 (Prph2,alsoknownas reduction of photoreceptor cells after 5 weeks. We conclude peripherin/RDS) is an abundantly expressed photoreceptor- that miRNA-based hairpins can achieve rapid and robust gene specific gene. In humans, gain-of-function mutations in PRPH2 silencing after efficient vector-mediated delivery to the retina. result in both autosomal dominant retinitis pigmentosa and The rationale of using an miRNA-based template to improve dominant maculopathies. Gene-silencing strategies for these the silencing efficiency of a hairpin may prove valuable for conditions include RNA interference by short hairpin RNAs allele-specific silencing in which the choice for an RNAi target is (shRNAs). Recent evidence suggests that microRNA (miRNA)- limited and offers an alternative therapeutic strategy for the based hairpins may offer a safer and more effective alternative. treatment of dominant retinopathies. In this study, we used for the first time a virally transferred Gene Therapy (2010) 17, 486–493; doi:10.1038/gt.2009.162; miRNA-based hairpin to silence Prph2 in the murine retina. published online 10 December 2009 Keywords: RNAi; Prph2; AAV; miRNA; rds; retinal degeneration Introduction retinal pigment epithelium.6 To date, over 90 different PRPH2 mutations that cause retinopathies have been Inherited retinal disorders are a major cause of blindness identified.7 Dominant retinopathies caused by mutations affecting 41.5 million individuals worldwide.1–5 Dom- in PRPH2 are relatively common, with PRPH2 mutations inantly inherited retinal degenerations are a subcategory involved in up to 25% of central retinal dystrophies, such of retinopathies that lead to progressive loss of photo- as a variety of macular dystrophies, cone-rod dystro- receptors caused either by haploinsufficiency because phies and central areolar choroidal dystrophies, as well of loss-of-function mutations, the deleterious effect of as 9% of adRP.7–11 Even though individuals with dominant negative mutations or by the toxic effect of dominant PRPH2 mutations often show inter- and gain-of-function mutations. To date, mutations in 45 intra-familial variation, most observe visual disturbances genes have been identified as a cause of dominant retinal by middle age. degenerations, including conditions such as autosomal Gene replacement therapy is a promising approach for dominant retinitis pigmentosa (adRP) and dominant the specific treatment of recessively inherited disorders maculopathies (Retnet Database, 2009). Peripherin-2 caused by loss-of-function mutations and a means to (Prph2) is a highly expressed photoreceptor-specific gene deliver neuroprotective proteins in retinal degenerations in which mutations often lead to dominantly inherited generically.5,12–14 However, therapies for dominant retinal disorders. The gene product, PRPH2, is a 32 kDa retinopathies, in which expression of a mutant allele structural membrane glycoprotein localized in the has a dominant negative effect on photoreceptor cell photoreceptor outer segments in which it forms multi- function or survival, are likely to depend on efficient subunit components with its homolog ROM-1. It is an silencing of the mutated allele. Ribozyme technology essential component in outer segment morphogenesis, a and RNAi have been used to target specifically mRNA vital daily process of photoreceptors to replenish the from mutated genes involved in adRP, and RNAi in outer segment tips that are being phagocytosed by the particular can result in efficient gene silencing.15–20 Local delivery of silencing short hairpin RNA (shRNA) cassettes using viral vectors such as AAV can result in Correspondence: Professor RR Ali, Department of Molecular sustained knockdown of the mutated gene, often leading Therapy, Institute of Ophthalmology, University College London, to prolonged improvement of the adRP phenotype.21–26 Bath Street 11-43, London EC1V 9EL, UK. E-mail: [email protected] Although vector-mediated RNAi offers a potentially Received 5 July 2009; revised 30 October 2009; accepted 31 October powerful approach to the treatment of dominant disease, 2009; published online 10 December 2009 the molecular kinetics of the RNAi pathway have yet to AAV-mediated knockdown of Prph2 in vivo A Georgiadis et al 487 be fully defined and the success in selecting a potent RNAi modulator still depends on a ‘hit-or-miss’ approach in which a range of RNAi molecules that span the coding sequence of the gene of interest are screened to identify the most efficient target area. Although such screening is appropriate where there is more than one potential target region, this method in unsuitable where there is no alternative, for instance in targeting a point mutation for allele-specific silencing. In addition to the fact that there are no effective solutions to date for the enhancement of an shRNA’s efficiency when choice of Figure 1 Generation of stable Prph2-expressing cell line and target area is limited, there is also a potential for siRNA knockdown. (a) Western blot for Prph2 on total protein significant off-target adverse effects.27,28 Solutions for lysates (20 mg) from 10 different clonal populations propagated from these problems in RNAi design may be found in single 293T cells transduced with a lentivirus-expressing Prph2. Clone #8 expressed Prph2 at approximately fivefold b-actin levels studying endogenous microRNAs (miRNAs). and was selected for subsequent experiments. (b) Three siRNAs- The discovery of endogenous miRNAs led to the targeting Prph2 at different regions (siRDS4,5,6) and a non-targeting development of ‘miRNA mimetics’ in which the design control (siCON) were transfected into 293T/RDS+ #8 at 20 mM. of miRNAs is used to accommodate RNAi target sequences to increase their silencing efficiency and, most importantly, reduce toxicity caused by shRNA-depen- dent saturation of the endogenous pathway. Indeed, miRNA-based design of RNAi hairpins has been shown to both increase efficiency and reduce toxicity possibly as a result of their intracellular processing.29–32 In this study, we evaluated the efficiency of an miRNA-based shRNA-targeting Prph2 (miRDS6) in vivo Figure 2 Nucleotide alterations between wtRDS and mutRDS. Site-directed PCR mutagenesis was performed in mutRDS to alter using AAV2/8 viruses. The efficiency of miRDS6 in nucleotides within the region targeted by RDS6. Three nucleotides silencing Prph2 was also compared with a conventionally (in bold italics) were altered using the degeneracy of the genetic designed hairpin (shRDS6). After subretinal delivery of code to retain codon characteristics. Both codon series encode for AVV2/8 pseudotyped vectors, histological analysis the amino-acids Tyrosine–Serine–Tyrosine. The relative abundances showed that the miRNA-based shRNA robustly and for each tRNA in the mouse are indicated and suggest that protein consistently silenced Prph2 expression both in vitro and translation would not be hampered by codon substitution. in vivo. This study provides the first assessment of a virally delivered miRNA in the retina for the silencing of an abundantly expressed photoreceptor gene and and such an engineered cDNA could be used for future shows the potential of miRNA mimetics in RNAi applications of suppression and replacement, which is applications for the treatment of dominantly inherited the process of gene supplementation with an engineered retinal degeneration. cDNA that is not silenced by the target hairpin after all endogenous expression has been abolished. Hence, the first three triplets within the first eight Results nucleotides or ‘seed region’ of RDS6 were altered to abolish shRNA:mRNA binding while retaining the wild- For the development of a potent Prph2-silencing hairpin, type amino-acid sequence (Figure 2). The mutated Prph2 three target sequences (siRDS4, siRDS5 and siRDS6) cDNA was named mutRDS and a stably expressing were selected from an online design website (http:// cell line (293T/mutRDS+) was constructed in a similar www.Dharmacon.com), together with a non-targeting manner as that used to generate 293T/wtRDS+. Having control (siCON), and their efficacy was assessed in vitro. established a potent target area for Prph2 knockdown We engineered a 293T-HEK cell line that stably expresses that was evaluated using siRNA transfections (siRDS6), Prph2 for hairpin evaluation. HEK cells were infected an miRNA-based hairpin with the same target area with a lentiviral vector expressing Prph2 from a spleen was designed. The miRNA template used was miR-30a34 focus-forming virus promoter at a limiting MOI (that is and the hairpin was named miRDS6. Similarly, the non- o1), and single cells were propagated to obtain a targeting control sequence (siCON) was also redesigned number of stable cell lines expressing Prph2 from a based on miR-30a (miRCON). In addition to miRCON single integrated expression cassette. The cell lines were and miRDS6, a conventional shRNA-targeting Prph2 characterized and Prph2 expression was quantified by at the same target area as miRDS6 was designed immunoblotting. We selected one cell line (293T/ (shRDS6) to be tested alongside miRDS6 in vitro and wtRDS+) in which Prph2 expression was fivefold higher in vivo (Figure 3). than endogenous b-actin expression (Figure 1a). This The two Prph2-targeting hairpin cassettes (shRDS6 cell line was used for assessing the three siRNAs.