SEROTYPES CAPSIDS AND SPECIFIC APPLICATIONS tissues. We speculate that AAV-1 vector may provide a valuable tool in therapeutic gene delivery for many diseases directly or indirectly involving endothelial cells.
95. Retargeting AAV-2 to Angiogenic Endothelial Cells with Tumor Vasculature Specific Capsid Variants Andreas Riederer,1 Bernard Kaess,1 Luca Perabo,1 Hildegard Buening,1 Michael Hallek.1,2,3 1 Gene Center, Ludwig Maximilian University of Munich, Munich, Germany; 2First Department of Internal Medicine, University Hospital Cologne, Cologne, Germany; 3GSF - National Centre for Research and Environment, KKG Gene Therapy, Munich, Germany. A promising target for cancer gene therapy is the tumor endothelium, which forms the inner wall of tumor vessels and is therefore essential for the oxygen and nutrient supply of the tumor. On these tumor vessels specific surface molecules are upregulated or specific isoforms are expressed. The development of targeting technologies allows the generation of viral vectors specifically tailored to bind to such molecules. Genetic incorporation of receptor specific ligands into the capsid position I-587 (Girod et al., 1999) seems to be the most promising approach for the generation of retargeting vectors based on the adeno-associated virus (AAV2). Therefore, we used this position to insert previously described tumor vasculature specific peptides into the capsid of AAV2 vectors. These targeting vectors could be generated with titers comparable to wild type AAV2. We could demonstrate that infections of these new retargeting vectors were mediated by the inserted peptides and that the same insertions interfered with the usage of AAV2s primary receptor heparan sulfate proteoglycan. Furthermore, two of the four peptides successfully shifted the viral tropism towards angiogenic HUVEC, an in vitro model for tumor vasculature. The herein described experiments show that it is possible to target native tumor vasculature at least in vitro with specific peptides inserted into the capsid of a viral vector.
96. Transgene Expression from Recombinant AAV-8 Vectors in the Liver Occurs Independently of the Serotype Origin of the Viral Inverted Terminal Repeats Dirk Grimm,1 Hiroyuki Nakai,1 Theresa A. Storm,1 Efren P. Riu,1 Mark A. Kay.1 1 Pediatrics, Stanford University School of Medicine, Stanford, CA. We present data from the long-term in vivo evaluation of gene transfer vectors combining properties from seven different serotypes of Adeno-Associated Virus (AAV), and discuss how they corroborate our current model for liver transduction by AAV. Amongst the eight different naturally occurring AAV serotypes developed as vectors, AAV-8 has emerged as the most potent candidate for liver transduction. Recently, we have provided evidence that rapid uncoating of the viral DNA from the AAV-8 capsid may be responsible for its enhanced transduction properties (Thomas et al., J. Virol. In press). Here, we have begun to analyse the role of another integral vector component, the inverted terminal repeats (ITRs), which are short sequences flanking the genome and traditionally derived from AAV serotype 2. ITRs are crucial for genome encapsidation, and also thought to be involved in viral DNA expression and persistence in the infected cell. To investigate how the origin of the ITRs affects the viral life cycle, we generated two series of AAV vector constructs carrying either the gfp, or the human factor IX (hfIX), gene flanked by ITRs from AAV serotypes 1 to 6. We initially addressed whether particular S38
ITRs, as well as serotype-specific proteins (Rep), would improve vector genome replication and encapsidation over an AAV-2-based standard approach. However, when cross-packaging each vector with each of six helper plasmids expressing AAV-1 to -6 proteins, we found no specific combination of ITRs and Rep advantageous. For subsequent animal studies, we thus used a common helper expressing AAV-2 Rep for large-scale packaging of the different hFIX constructs into AAV-8 virions. An exception was the vector carrying AAV-5 ITRs, which was exclusively processed by AAV-5 Rep, thus forcing us to develop a novel hybrid helper expressing those proteins next to AAV-8 capsid proteins. The six hFIX vectors were next intraportally injected into immunocompetent mice, each at three increasing doses of vector genomes, low (5e10), middle (3e11), and high (1.8e12), and plasma hFIX levels were determined by ELISA over a six month period. Two findings were most remarkable : 1) there was no significant difference in hFIX levels expressed at a given particle dose between the 6 vectors, and 2) expression peaked between week 2 and 4 after injection, with individual mice showing hFIX levels up to 1 mg/ml, before declining to stable levels similar to those seen in the first week. To corroborate the identical protein expression profiles on the DNA level, we performed partial hepatectomies on selected mice to study the molecular state of the proviral genomes in transduced hepatocytes. The results of this ongoing process will be presented. We conclude that in the liver, AAV ITRs are interchangeable, without affecting short- or long-term transduction profiles. This could indirectly support our current model of AAV liver transduction, according to which AAV-mediated gene expression requires annealing of single-stranded vector genomes, a process that is presumably ITR-independent. Our findings also provide a rationale for developing synthetic ITRs which could enhance AAV vector replication and encapsidation, ultimately improving particle production.
97. Selection of Potent AAV Pseudotypes To Transduce Rat Pancreatic Islets Brigitte Le Mauff, Françoise Boeffard, Nasro Benrezzak, Ignacio Anegon. 1 U643, INSERM/ITERT, Nantes, France. We have already shown that CTLA4Ig expression in rat islet grafts, using an adenoviral vector (AdCTLA4Ig) significantly prolonged allograft survival, albeit less efficiently than systemic production of CTLA4Ig following intra-muscular injection of AdCTLA4Ig. Maximum transduction by the adenoviral vector could not be achieved in vivo due to a high level of primary non graft function when using MOI > 102 infectious particles (ip)/cell (105 ip/ islet). To overcome this limitation, we compared various pseudotypes of AAV vectors, AAV1- 2- 4- 5- 8, all containing EGFP under the transcriptional control of the CMV promoter. The process of production was completely free of adenovirus and resulted in viral titers from 5.1010 to 2.1013 particles/ml according to the serotypes. After isolation, handpicked islets were incubated with 3. 104 particles/cell, without adenovirus co-infection, and maintained in culture for GFP detection using microscopy and measurement with a fluorometer. AAV1 was constantly the most potent vector, followed by AAV4, but the latter was toxic at this dose. AAV5 and 8 were also quite efficient in contrast to AAV2 which was previously the most commonly used AAV vector. Significant levels of GFP were observed at day 3 but expression still increased and plateaued around d30. Although levels of transduction were lower than those obtained with the Ad vector used at 10 ip/cell, islet viability was better conserved with AAV (except AAV4) than with adenovirus where islets were completely dead between d20 and d30. Various AAV pseudotype vectors encoding CTLA4Ig have been generated Molecular Therapy Volume 9, Supplement 1, May 2004
Copyright The American Society of Gene Therapy