Beverly L. Davidson

RNA polymerase III transcribes human microRNAs

Prior work demonstrates that mammalian microRNA (miRNA or miR) expression requires RNA polymerase II (Pol II). However, the transcriptional requirements of many miRNAs remain untested. Our genomic analysis of miRNAs in the human chromosome 19 miRNA cluster (C19MC) revealed that they are interspersed among Alu repeats. Because Alu transcription occurs through RNA Pol III recruitment, and we found that Alu elements upstream of C19MC miRNAs retain sequences important for Pol III activity, we tested the promoter requirements of C19MC miRNAs. Chromatin immunoprecipitation and cell-free transcription assays showed that Pol III, but not Pol II, is associated with miRNA genomic sequence and sufficient for transcription. Moreover, the mature miRNA sequences of approximately 50 additional human miRNAs lie within Alu and other known repetitive elements. These findings extend the current view of miRNA origins and the transcriptional machinery driving their expression.

siRNA-mediated gene silencing in vitro and in vivo

RNA interference is now established as an important biological strategy for gene silencing, but its application to mammalian cells has been limited by nonspecific inhibitory effects of long dsRNA on translation. Here, we describe a viral-mediated delivery mechanism that results in specific silencing of targeted genes through expression of small interfering RNA (siRNA). We establish proof of principle by markedly diminishing expression of exogenous and endogenous genes in vitro and in vivo in brain and liver, and further apply this strategy to a model system of a major class of neurodegenerative disorders, the polyglutamine diseases, to show reduced polyglutamine aggregation in cells. This viral-mediated strategy should prove generally useful in reducing expression of target genes to model biological processes or to provide therapy for dominant human diseases.

Transvascular delivery of small interfering RNA to the central nervous system

A major impediment in the treatment of neurological diseases is the presence of the blood–brain barrier, which precludes the entry of therapeutic molecules from blood to brain. Here we show that a short peptide derived from rabies virus glycoprotein (RVG) enables the transvascular delivery of small interfering RNA (siRNA) to the brain. This 29-amino-acid peptide specifically binds to the acetylcholine receptor expressed by neuronal cells. To enable siRNA binding, a chimaeric peptide was synthesized by adding nonamer arginine residues at the carboxy terminus of RVG. This RVG-9R peptide was able to bind and transduce siRNA to neuronal cells in vitro, resulting in efficient gene silencing. After intravenous injection into mice, RVG-9R delivered siRNA to the neuronal cells, resulting in specific gene silencing within the brain. Furthermore, intravenous treatment with RVG-9R-bound antiviral siRNA afforded robust protection against fatal viral encephalitis in mice. Repeated administration of RVG-9R-bound siRNA did not induce inflammatory cytokines or anti-peptide antibodies. Thus, RVG-9R provides a safe and noninvasive approach for the delivery of siRNA and potentially other therapeutic molecules across the blood–brain barrier.

Superoxide Production in Vascular Smooth Muscle Contributes to Oxidative Stress and Impaired Relaxation in Atherosclerosis

The endothelium is a source of reactive oxygen species in short-term models of hypercholesterolemia and atherosclerosis. We examined a chronic model of atherosclerosis for increased vascular production of superoxide (O2-.) and determined whether endothelial overexpression of superoxide dismutase (SOD) would improve endothelium-dependent relaxation. Superoxide generation was 3 times higher in isolated aortas from Watanabe heritable hyperlipidemic (WHHL) rabbits (2 to 4 years old) compared with aortas from New Zealand White (NZ) rabbits (43+/-10 versus 14+/-2 relative light units x min(-1) x mm(-2), n=9, P<0.05). After in vitro transduction with adenovirus containing the gene for CuZn-SOD (AdCMVCuZn-SOD) or extracellular SOD (AdCMVEC-SOD), endothelial O2-. levels in WHHL aortas were significantly reduced. Gene transfer of SOD to WHHL aortas, however, failed to improve the impaired relaxation to acetylcholine or calcium ionophore. By use of the oxidative fluorescent dye hydroethidine, an in situ assay indicated markedly increased generation of O2-. throughout the wall of WHHL aorta, especially within layers of smooth muscle. This finding was confirmed by demonstrating increased O2-. levels in smooth muscle cells cultured from WHHL aorta. We conclude that elevated O2-. levels in atherosclerotic vessels are not confined to the endothelium but occur throughout the vascular wall, including smooth muscle cells. Reduction in endothelial O2-. levels is not sufficient to improve endothelium-dependent relaxation. Generation of reactive oxygen species within the media may contribute to vasomotor dysfunction in atherosclerosis.

RNAi suppresses polyglutamine-induced neurodegeneration in a model of spinocerebellar ataxia

The dominant polyglutamine expansion diseases, which include spinocerebellar ataxia type 1 (SCA1) and Huntington disease, are progressive, untreatable, neurodegenerative disorders. In inducible mouse models of SCA1 and Huntington disease, repression of mutant allele expression improves disease phenotypes. Thus, therapies designed to inhibit expression of the mutant gene would be beneficial. Here we evaluate the ability of RNA interference (RNAi) to inhibit polyglutamine-induced neurodegeneration caused by mutant ataxin-1 in a mouse model of SCA1. Upon intracerebellar injection, recombinant adeno-associated virus (AAV) vectors expressing short hairpin RNAs profoundly improved motor coordination, restored cerebellar morphology and resolved characteristic ataxin-1 inclusions in Purkinje cells of SCA1 mice. Our data demonstrate in vivo the potential use of RNAi as therapy for dominant neurodegenerative disease.

Recombinant adeno-associated virus type 2, 4, and 5 vectors: Transduction of variant cell types and regions in the mammalian central nervous system

Recombinant adeno-associated virus vectors based on serotype 2 (rAAV2) can direct transgene expression in the central nervous system (CNS), but it is not known how other rAAV serotypes perform as CNS gene transfer vectors. Serotypes 4 and 5 are distinct from rAAV2 and from each other in their capsid regions, suggesting that they may direct binding and entry into different cell types. In this study, we examined the tropisms and transduction efficiencies of beta-galactosidase-encoding vectors made from rAAV4 and rAAV5 compared with similarly designed rAAV2-based vectors. Injection of rAAV5 beta-galactosidase (betagal) or rAAV4betagal into the lateral ventricle resulted in stable transduction of ependymal cells, with approximately 10-fold more positive cells than in mice injected with rAAV2betagal. Major differences between the three vectors were revealed upon striatal injections. Intrastriatal injection of rAAV4betagal resulted again in striking ependyma-specific expression of transgene, with a notable absence of transduced cells in the parenchyma. rAAV2betagal and rAAV5betagal intrastriatal injections led to beta-gal-positive parenchymal cells, but, unlike rAAV2betagal, rAAV5betagal transduced both neurons and astrocytes. The number of transgene-positive cells in rAAV5betagal-injected brains was 130 and 5,000 times higher than in rAAV2betagal-injected brains at 3 and 15 wk, respectively. Moreover, transgene-positive cells were widely dispersed throughout the injected hemisphere in rAAV5betagal-transduced animals. Together, our data provide in vivo support for earlier in vitro work, suggesting that rAAV4 and rAAV5 gain cell entry by means of receptors distinct from rAAV2. These differences could be exploited to improve gene therapy for CNS disorders.

Current prospects for RNA interference-based therapies

RNA interference (RNAi) is a powerful approach for reducing expression of endogenously expressed proteins. It is widely used for biological applications and is being harnessed to silence mRNAs encoding pathogenic proteins for therapy. Various methods — including delivering RNA oligonucleotides and expressing RNAi triggers from viral vectors — have been developed for successful RNAi in cell culture and in vivo. Recently, RNAi-based gene silencing approaches have been demonstrated in humans, and ongoing clinical trials hold promise for treating fatal disorders or providing alternatives to traditional small molecule therapies. Here we describe the broad range of approaches to achieve targeted gene silencing for therapy, discuss important considerations when developing RNAi triggers for use in humans, and review the current status of clinical trials.

Artificial miRNAs mitigate shRNA-mediated toxicity in the brain : Implications for the therapeutic development of RNAi

Huntingtons disease (HD) is a fatal, dominant neurodegenerative disease caused by a polyglutamine repeat expansion in exon 1 of the HD gene, which encodes the huntingtin protein. We and others have shown that RNAi is a candidate therapy for HD because expression of inhibitory RNAs targeting mutant human HD transgenes improved neuropathology and behavioral deficits in HD mouse models. Here, we developed shRNAs targeting conserved sequences in human HD and mouse HD homolog (HDh) mRNAs to initiate preclinical testing in a knockin mouse model of HD. We screened 35 shRNAs in vitro and subsequently narrowed our focus to three candidates for in vivo testing. Unexpectedly, two active shRNAs induced significant neurotoxicity in mouse striatum, although HDh mRNA expression was reduced to similar levels by all three. Additionally, a control shRNA containing mismatches also induced toxicity, although it did not reduce HDh mRNA expression. Interestingly, the toxic shRNAs generated higher antisense RNA levels, compared with the nontoxic shRNA. These results demonstrate that the robust levels of antisense RNAs emerging from shRNA expression systems can be problematic in the mouse brain. Importantly, when sequences that were toxic in the context of shRNAs were placed into artificial microRNA (miRNA) expression systems, molecular and neuropathological readouts of neurotoxicity were significantly attenuated without compromising mouse HDh silencing efficacy. Thus, miRNA-based approaches may provide more appropriate biological tools for expressing inhibitory RNAs in the brain, the implications of which are crucial to the development of RNAi for both basic biological and therapeutic applications.

Allele-specific silencing of dominant disease genes

Small interfering RNA (siRNA) holds therapeutic promise for silencing dominantly acting disease genes, particularly if mutant alleles can be targeted selectively. In mammalian cell models we demonstrate that allele-specific silencing of disease genes with siRNA can be achieved by targeting either a linked single-nucleotide polymorphism (SNP) or the disease mutation directly. For a polyglutamine neurodegenerative disorder in which we first determined that selective targeting of the disease-causing CAG repeat is not possible, we took advantage of an associated SNP to generate siRNA that exclusively silenced the mutant Machado–Joseph disease/spinocerebellar ataxia type 3 allele while sparing expression of the WT allele. Allele-specific suppression was accomplished with all three approaches currently used to deliver siRNA: in vitro-synthesized duplexes as well as plasmid and viral expression of short hairpin RNA. We further optimized siRNA to specifically target a missense Tau mutation, V337M, that causes frontotemporal dementia. These studies establish that siRNA can be engineered to silence disease genes differing by a single nucleotide and highlight a key role for SNPs in extending the utility of siRNA in dominantly inherited disorders.

Adeno-Associated Virus Type 5 (AAV5) but Not AAV2 Binds to the Apical Surfaces of Airway Epithelia and Facilitates Gene Transfer

ABSTRACT In the genetic disease cystic fibrosis, recombinant adeno-associated virus type 2 (AAV2) is being investigated as a vector to transfer CFTR cDNA to airway epithelia. However, earlier work has shown that the apical surface of human airway epithelia is resistant to infection by AAV2, presumably as a result of a lack of heparan sulfate proteoglycans on the apical surface. This inefficiency can be overcome by increasing the amount of vector or by increasing the incubation time. However, these interventions are not very practical for translation into a therapeutic airway-directed vector. Therefore, we examined the efficiency of other AAV serotypes at infecting human airway epithelia. When applied at low multiplicity of infection to the apical surface of differentiated airway epithelia we found that a recombinant AAV5 bound and mediated gene transfer 50-fold more efficiently than AAV2. Furthermore, in contrast to AAV2, AAV5-mediated gene transfer was not inhibited by soluble heparin. Recombinant AAV5 was also more efficient than AAV2 in transferring β-galactosidase cDNA to murine airway and alveolar epithelia in vivo. These data suggest that AAV5-derived vectors bind and mediate gene transfer to human and murine airway epithelia, and the tropism of AAV5 may be useful to target cells that are not permissive for AAV2.