371. Complete Inhibition of Hepatitis B Virus Gene Expression In Vivo with Short Hairpin RNA Expressed from a Novel Double-Stranded, Bi-Cistronic Adeno-Associated Virus Pseudotype 8 Vector

Abstract

Top of pageAbstract We report the generation and in vivo evaluation of novel Adeno-Associated Viral (AAV) vectors for the expression of short hairpin RNA (shRNA) against human Hepatitis B Virus (HBV), and demonstrate their superior efficiency over conventional AAV vectors. Our approach combines two recent major advances in the AAV field: pseudotyping of AAV-2-based vector genomes with capsids from another serotype, and development of “double-stranded” (ds) vectors, that express faster and more strongly than standard, single-stranded (ss) recombinant AAV. In detail, we cross-packaged AAV genomes comprising our most robust anti-HBV shRNA, or a control shRNA against luciferase, into capsids from AAV-8, which is the AAV serotype with the highest known transduction rate in liver. To force their encapsidation as ds genomes, we replaced one AAV-2 packaging signal with a mutant derived from AAV-4 or -5, which concurrently improved vector genome stability and particle titers. The small size of the shRNA inserts enabled us to encapsidate a second cistron, encoding a fluorescence protein for easy detection of shRNA-expressing cells. Specificity and efficiency of all vectors was first analysed in a human HBV-transformed cell line. As hoped for, only the two anti-HBV vectors significantly reduced HBV surface antigen (sAg) expression. Importantly, the novel dsAAV showed a much stronger effect, and more rapidly, than the standard ssAAV. All vectors were next evaluated in transgenic mice carrying and expressing an integrated HBV genome. Particles were given intraportally or by peripheral intravenous infusion, at doses from 5e10 up to 7e12 vector genomes (vg) per animal. Most remarkably, analyses of serum before and after systemic delivery of the ds anti-HBV vector showed complete elimination of sAg between day 4 and 11 after injection. This was found at doses as low as 3e11 vg per animal, whereas to get the same effect with the ss vector, 20-fold higher doses and longer incubation were required. No significant change in sAg was seen with the anti-luciferase control vectors or after PBS injection. Those results were corroborated by HBV mRNA and DNA analyses, and were confirmed in a second lineage of HBV-transgenic mice. Interestingly, we noted a dose-dependent toxicity of the anti-HBV vectors, resulting in hypoproteinemia and ultimately death of animals given high particle doses (1.8e12 vg and greater). This was seen as early as 5 days after injection, and in both strains of HBV-transgenic as well as in normal mice, ruling out a role for the integrated HBV target genome. The fact that toxicity was absent from the control shRNA vectors indicated an adverse off-target effect specific for the anti-HBV shRNA used here, which is a current matter of further investigation. In sum, our novel dsAAV-8 vector system provides an utmost efficient means for delivery of shRNA to the liver, allowing the use of minimal vector doses and non-invasive routes of administration. This makes it highly interesting for further development towards clinical exploitation of RNA interference to treat hepatitis and other liver diseases.