AAV VECTORS: VECTOR TRAFFICKING 6. Novel Cardiac-Tropic AAV Vectors by DNA Shufing and In Vivo Screening
Zhenhua Yuan,1 Lin Yang,1 Chunping Qiao,1 Juan Li,1 Xiao Xiao.1 1 Division of Molecular Pharmaceutics, Eshelman School of Pharmacy, University of North Carolina, Chapel Hill, NC. AAV-mediated gene therapy is one of the most promising methods for treatment of heart disease. It is highly desirable to efciently deliver therapeutic genes into the heart but not unintended tissues. To address this issue, generation of novel cardiac tropic AAVs via direct in vivo screening from AAV capsid library is considered as a feasible and promising approach. Here we report the novel heart tropic AAVs, AAVH50 and AAVH59 generated by DNA shufing and direct in vivo screening. Methods: Firstly, we established an AAV cap gene library by DNA shufing of different natural AAV serotype capsid genes. Secondly, the infectious AAV library with shufed capsid genes was packaged and then directly screened in mice in vivo. In brief, the AAV capsid library was originated from DNA shufing of different natural AAV serotypes (AAV1, 2, 3, 4, 6, 7, 8 and 9). The infectious AAV library was intravenously injected into C57BL/6 mice, and different mouse tissues were collected for AAV capsid gene retrieval and sequencing. The clones enriched in heart but scarce in the liver was obtained and subjected to further characterization after two-round of in vivo screening in mice. Results: Two clones designated as AAVH50 and AAVH59 were retrieved mainly based on its high frequency in the heart and low frequency in the liver. We then compared the gene transduction efciency as well as vector distributions of those two novel vectors along with AAV6, a parental AAV with strong muscle tropism, and AAV9, the most robust AAV for systemic gene delivery using luciferase reporter gene in C57BL/6 mice. We found that AAVH50 was more efcient than AAV6, and was very close to AAV9 in the heart, but less efcient than AAV9 (> 30-fold lower) and AAV6 (> 25-fold lower) in the liver. AAVH50 had the highest gene expression in the myocardium among all tissues tested. Strikingly, AAVH59 showed barely above background level expression in the liver, but only slight lower expression in heart and higher expression in muscle than AAVH50. Quantitative PCR showed that the vector DNA copy numbers in the liver of H50-treated mice was 169-fold lower than that of AAV9, 30-fold lower than that of AAV6. AAVH59 liver copy number was 55000-fold lower than that of AAV9, and 9830-fold lower than that of AAV6, consistently with luciferase expression data. Moreover, similar results were also obtained using a different strain of mouse (ICR ) or a different reporter gene (lacZ gene). Finally, we analyzed the direct transduction efciency of AAVH50 and AAVH59 on primary rat cardiomycytes culture. The result displayed that the infectivity of AAVH50 and AAVH59 for cardiomyocytes was higher than AAV9 but lower than AAV6. Conclusion: Our preliminary data demonstrated that direct evolution of AAV via DNA shufing and direct in vivo screen is a powerful and reliable strategy to obtain novel de-targeting and re-targeting AAV vector. AAVH50 and AAVH59 could be useful in pre-clinical animal experiments and potentially clinical applications for gene therapy of cardiac diseases.
7. Intravenously Administered rAAV9 Vector Efciently Transduces Tissues In Vivo Via Caveolin-1-Independent Transport across the Capillary Endothelial Cell Barrier
Kei Adachi,1 Nicole M. Kotchey,1 Maliha Zahid,1 Hiroyuki Nakai.1 1 Department of Microbiology and Molecular Genetics, University of Pittsburgh School of Medicine, Pittsburgh, PA. Recombinant AAV serotype 9 (rAAV9) vectors transduce various tissues robustly following systemic administration in vivo. The tissues most permissive to rAAV9 are the liver and heart, in which the anatomical structures of vascular endothelium are signicantly different. Hypothetically, circulating rAAV particles can have direct Molecular Therapy Volume 18, Supplement 1, May 2010 Copyright © The American Society of Gene & Cell Therapy
access to hepatocytes via fenestrae in liver sinusoidal endothelial cells, which explains preferential transduction in the liver with many rAAV serotypes. In non-hepatic tissues including the heart, however, capillary endothelium forms a signicant physical and biological barrier precluding efcient parenchymal cell transduction with many rAAV serotypes administered via intravascular routes. Since rAAV9 efciently transduces non-hepatic parenchymal cells via the blood stream, it has been proposed that rAAV9 efciently crosses capillary endothelial cell barriers. Endothelial cells are permeable to various macromolecules, and caveolae-mediated transcytosis is the major macromolecule transport pathway across the endothelial cell barrier. Although signicantly enhanced transendothelial transport of rAAV9 via transcytosis represents a potential mechanism, the exact mechanisms of rAAV9’s robustness in systemic transduction remain largely unknown. Here, we show that transendothelial transport of rAAV9 in vivo is a slow process, has limited capacity, and occurs in the absence of caveolin-1, the major component of caveolae that mediate transcytosis. In our studies, we intravenously infused rAAV9 or rAAV1.9-6 (a rAAV variant) into mice, and altered vector exposure time by injecting the corresponding neutralizing antibodies that rapidly eliminate circulating viral particles. rAAV9 and rAAV1.9-6 are recombinant vectors that transduce the heart robustly but show contrasting blood clearance rates (i.e., rAAV9 exhibits a very slow blood clearance while rAAV1.9-6 is rapidly cleared from the blood circulation). By analyzing pharmacokinetic profiles and vector transduction efciency in the liver, heart and other tissues, we nd that transendothelial transport of rAAV in non-hepatic tissues takes 1-4 hours, while a large amount of rAAV particles freely and rapidly cross liver sinusoidal endothelial cells, completing nearly full hepatic parenchymal cell infection in 4 hours. The presence and absence of the lag between endothelial exposure to rAAV9 or rAAV1.9-6 and their infection to parenchymal cells in the non-hepatic and hepatic tissues, respectively, suggests signicant involvement of a transendothelial viral transport mechanism in non-hepatic tissue transduction with rAAV. This process, however, is not at all affected in caveolin-1 knockout mice devoid of caveolae-mediated transcytosis. Thus, our observations demonstrate that capillary endothelium remains a signicant barrier to rAAV9 and support an alternative mechanism; i.e., the distinctively slowed blood clearance of rAAV9 sustains slow and capacity-limited caveolin-1-independent transendothelial vector transport to parenchymal cells, resulting in robust non-hepatic tissue transduction with rAAV9.
8. Directed Evolution of Adeno-Associated Virus Variants That Evade Human Neutralizing Antibodies
Daniel Stone,1 James T. Koerber,1 Federico Mingozzi,3 Gregory M. Podsakoff,3 Katherine High,3,4 David V. Schaffer.1,2 1 Department of Chemical Engineering, University of California, Berkeley, CA; 2Helen Wills Neuroscience Institute, University of California, Berkeley, CA; 3Division of Hematology and Center for Cellular and Molecular Therapeutics, The Children’s Hospital of Philadelphia, Philadelphia, PA; 4Howard Hughes Medical Institute, The Children’s Hospital of Philadelphia, Philadelphia, PA. Despite the recent success of adeno-associated virus (AAV) vectors in several clinical trials, host-mediated immune responses remain a signicant hurdle to their future use. For example, in patients who have been pre-exposed to wild type AAV, or have already received AAV as a therapeutic, the development of anti-virus neutralizing antibodies (NAbs) that inhibit transduction is a serious hindrance to gene delivery. We have generated a series of AAV capsid variant libraries and applied an in vitro infection/rescue process to select for mutants that evade human NAbs. Several libraries were generated and utilized: a randomly mutagenized AAV2 library, an AAV2 S3