303. Utilizing DNA Shuffling Technologies To Generate AAV Libraries To Select for and Evolve Capsids With an Expanded Packaging Capacity

303. Utilizing DNA Shuffling Technologies To Generate AAV Libraries To Select for and Evolve Capsids With an Expanded Packaging Capacity

AAV VECTOR DEVELOPMENT at the same time points. Interestingly, the presence of anti-AAV8 antibodies did not interfere with persistence of luciferase e...

239KB Sizes 0 Downloads 2 Views

AAV VECTOR DEVELOPMENT at the same time points. Interestingly, the presence of anti-AAV8 antibodies did not interfere with persistence of luciferase expression. The lack of detectable anti-AAVHSC15 antibodies at 4 weeks postinjection suggested that this capsid may be less immunogenic than AAV8. It is also possible that the differences in transgene expression between immunodeficient and immunocompetent mice may stem from differences in DNA-dependent protein kinase activity, which is mutated in NOD/SCID mice. Overall, we conclude that while both AAV8 and AAVHSC15 are excellent candidates for intramuscular transgene expression, AAVHSC15 supports higher levels of expression than AAV8, particularly in NOD/SCID mice. This, combined with the apparent lower immunogenicity of AAVHSC15, makes it an attractive gene delivery vector for the muscle.

301. An Efficient Production Method for Dual Recombinant AAV Vectors

Qizhao Wang,1 Biao Dong,1 Sean Roberts,1 Jenni Firrman,1 Yong Diao,1 Wenjing Cao,2 Ruian Xu,2 Weidong Xiao.1 1 Department of Microbiology and Immunology, Sol Sherry Thrombosis Research Center, Temple University, Philadelphia, PA; 2 School of Biomedical Sciences, Huaqiao University, Quanzhou, Fujian, China. Recombinant adeno-associated viral (rAAV) vectors have gained attention for gene therapy because of their safety and clinical efficacy. It remains a viable strategy to split the factor VIII gene into two AAV vectors to avoid the limitation of their small packaging capacity (~5.0kb). However, it is time-consuming and labor intensive to produce two rAAV vectors in separate batches for targeting the liver. Here we demonstrated that dual rAAV vectors could be successfully produced in a single preparation process using a dual vector plasmid transfection method. When the AAV vector plasmids carrying the human factor VIII heavy chain (hHC) and the light chain (hLC) expression cassettes were co-transfected into 293 cells along with the AAV-rep&cap and mini-adenovirus plasmids, both rAAV-hHC and rAAV-hLC were produced at the desired ratio and in high titer. Interestingly, the rAAV-hHC vectors always yielded higher titers than rAAV-hLC vectors due to more efficiently replicating rAAV-hHC genomes. There were no unexpected recombination events observed in this method of rAAV vector production. The vectors produced by this method were efficient in transducing the tissue culture cells in vitro. When these vectors were administered to hemophilia A mice, functional factor VIII was detected in the mouse plasma by both aPTT and ELISA assays. The functional units as well as the antigen levels of factor VIII were similar to those vectors produced by the traditional method and mixed at the similar ratio, both in vitro and in vivo. The dual vector plasmid method has been successfully used to produce rAAV vectors based on both AAV2 and AAV8 serotypes. This suggests that the mixed vector DNA method for producing rAAV vectors is serotype independent. In conclusion, the dual vector DNA strategy for AAV vector production is an effective method that may reduce the cost and labor for producing large quantity of rAAV vectors.

302. Hybrid Adeno-Associate Virus Vectors with Recombinant Capsids Improve Viral Production Yield and Gene Delivery Efficiency

Chuansong Wang,1 En-Ju Lin,1 Colin Stets,1 Alexandre Mouravlev,2 Meng Sun,1 Deborah Young,2 Matthew J. During.1,2 1 Department of Melcular Virology, Immunology and Medical Genetics; and Comprehensive Cancer Center, The Ohio State University, Columbus, OH; 2Department of Molecular Medicine and Pathology, The University of Auckland, Auckland, New Zealand. PURPOSES: Adeno-associated viral (AAV) vectors have shown promise as therapeutic agents for neurologic disorders. Thus, improving the viral production yield and gene delivery efficiency has become important to increase the success of basic research and clinical trials. In this study, 4 novel hybrid AAV vectors (named AAVRec 1 – 4) with recombinant capsids (structural domains of different rAAV serotypes isolated from primate brain were combined) were compared to other natural AAV variants (e.g. AAV1, AAV8, AAV9) for viral production yield and gene delivery efficiency following mice brain injection. METHODS: An expression cassette containing the CAG promoter driving the green fluorescent protein (GFP) gene is flanked by AAV2 inverted terminal repeats (ITR). The helper plasmids for each serotype are AAVRec1-4, AAV1, AAV8 and AAV9. The cis plasmid, AAV helper plasmid and Adeno-helper plasmid were cotransfected by CaPO4 in HEK 293 cells. The iodixanol density ultracentrifugation method was used to purify virus. Each serotype virus was produced in five 150 mm plates. Virus genomic titer of each vector stock from each plate was determined by real-time PCR, and virus yield (virus genomic particles per cell, vg/cell) in each plate calculated. To compare gene delivery efficiency, 2μl of 1 x 1012 vg/ ml AAV/GFP vectors (AAVRec1, AAVRec2, AAVRec3, AAVRec4, AAV1, AAV8 and AAV9) were injected into the striatum unilaterally via stereotaxic neurosurgery. Four weeks after vector injection, mice were sacrificed. The brains were collected for evaluation of GFP expression by unbiased stereological analysis of the GFP fluorescence. RESULTS: AAVRec 1, 2 and 3 viral yield (1.3±0.5, 1.7±0.8, 1.2±0.4 x 105 vg/cell respectively) are significantly higher (P< 0.05) than AAV1, 8 and 9 (0.4±0.3, 0.6±0.3, 0.1±0.1 x 105 vg/cell respectively). AAVRec4 viral yield (0.6±0.3 x 105 vg/cell) is similar with AAV8. GFP expression comparison, the increased GFP fluorescence intensity and spread for AAVRec3 and AAVRec4 were observed in the striatum. CONCLUSIONS: Hybrid AAV Vectors (AAVRec1 – 3) have higher viral production yield than AAV1, 8 and 9. AAVRec3 and AAVRec4 have higher gene delivery efficiency than AAV1 and AAV8 in the brain. These new hybrid AAV vectors may have high potential for future gene therapy for neurological disorders.

303. Utilizing DNA Shuffling Technologies To Generate AAV Libraries To Select for and Evolve Capsids With an Expanded Packaging Capacity Nicole Paulk,1 Katja Pekrun,1 Mark Kay.1 1 Human Gene Therapy, Stanford University, Stanford, CA.

A significant limitation for gene therapists is the small packaging capacity of AAV that places a limit on the number of therapeutic genes that can be delivered, and thus the variety of treatable diseases. Our laboratory has utilized a replicating AAV capsid library to screen for and direct the evolution of AAV capsids with the ability to package a larger genome to better meet clinical needs. Importantly, the use of a replicating AAV library selects and evolves for parameters in addition to receptor binding and cellular uptake such as capsid uncoating, which our lab has demonstrated to be important in defining transduction efficiencies between species and tissues. Here, actively replicating AAV libraries generated by DNA shuffling were used to screen for capsids with the ability to package S116

Molecular Therapy Volume 22, Supplement 1, May 2014 Copyright © The American Society of Gene & Cell Therapy

AAV VECTOR DEVELOPMENT one of several large genomes in human cells. As a control, we demonstrated that modifications to expand the AAV genome did not affect the ability to package a standard-sized genome. Different insert sizes (100bp, 500bp, 1kb and 2kb) were cloned into the noncoding region of the AAV genome downstream of CAP and prior to the ITR. Each of the four shuttle plasmids were vectorized and tested to confirm that they could still package a routine-sized genome (rAAV-ApoE/hAAT-eGFP, 4.7kb) as efficiently as AAV2. Controlverified plasmid libraries were used to produce large preparations of actively replicating AAV libraries. Huh7 cells were infected with AAV libraries at a range of MOIs and subsequently with adenovirus for AAV replication. Presence of adenovirus induces cell lysis and release of AAVs into the media. Media is collected 3 days later and used for western blot analysis of VP1/2/3. To avoid cross-packaging, the sample with the lowest detectable level of VP is used for the next round of selection. Selective pressure was applied for several screen rounds and the resultant capsid variants were fully Sanger sequenced and vectorized. Capsids are now being used to generate large capacity AAV-GFP vectors. Each lot will be titered by dot blot and those that produce virus at <1e10 vg/mL will not be considered further for use as a clinical candidate. Alkaline Southern blots are also being performed to verify an improvement in packaged genome size pre- and post-transduction. Cellular transduction efficiencies will be assessed on a variety of human cell lines from numerous organs and liver cancer models. The use of AAV capsid libraries may provide a way to overcome packaging size limitations inherent to current AAV vectors.

304. The Inhibitory Effect of Various Transfection Agents on rAAV Vector-Mediated Transgene Expression Both In Vitro and In Vivo

Yuan Wang,1,2 Lina Wang,1 George V. Aslanidi,3,4 Changquan Ling,1,2 Arun Srivastava,3,4 Chen Ling.3,4 1 Department of Traditional Chinese Medicine, Second Military Medical University, Shanghai, China; 2Shanghai University of Traditional Chinese Medicine, Shanghai, China; 3Division of Cellular and Molecular Therapy, Department of Pediatrics, College of Medicine, University of Florida, Gainesville, FL. The ability of transfection agents (TAs) to introduce DNA into cells has provided a powerful means to study the function and control of mammalian genes, and in some instances to examine the roles of specific protein(s) in recombinant adeno-associated viral (rAAV) vector-mediated transduction. A wide variety of TAs are available, including cationic polymers, cationic lipids, dendrimers, and calcium phosphate-base agents, all of which have been introduced to study various steps in the life cycle of rAAV vectors. For example, we evaluated the role of a cellular serine/threonine protein phosphatase, protein phosphatase 5 (PP5), in rAAV2 vector-mediated gene transfer using liposome-based lipids (Gene Ther., 14(6): 545-550, 2007). Similarly, Nonnenmacher et al. (Cell Host Microbe, 10(6): 563-576, 2011) utilized cationic polymers to deliver various genes into HeLa cells and concluded that rAAV2 vectors use the pleiomorphic CLIC/ GEEC pathway as the major endocytic infection route. Liu et al. (Gene Ther., 20(3): 308-317, 2013), on the other hand, transfected HEK293 cells with DsRed-Rab5, -Rab7 or -Rab11 expression plasmids by the calcium phosphate precipitation method and concluded that late endosomes might not be involved in the rAAV8 transduction. Yet, there have been no systematic studies comparing the effects of various TAs on rAAV vector-mediated transgene expression. Here, we report that, in the absence of recombinant plasmids, all TAs, inhibited rAAV2 vector transduction efficiency in various cell lines in vitro. For example, non-liposomal lipid formulations inhibited the transduction efficiency by >1.8-fold, lipid-based liposomes by >1.2fold, and calcium phosphate-based agents by >3-fold. Interestingly, cationic polymers and dendrimers inhibited viral-mediated transgene Molecular Therapy Volume 22, Supplement 1, May 2014 Copyright © The American Society of Gene & Cell Therapy

expression by >10-fold. Similar results were obtained in the presence of plasmids, except that non-liposomal lipids inhibited the transduction efficiency by >4-fold. In addition, we observed that the inhibitory effect of the TAs was largely limited to pre-treatment of cells, or treatment within the first 24 hours of viral transduction, suggesting that the intracellular trafficking step of vector particles was being affected. We also compared the effects of cationic polymers and lipid-based liposomes on rAAV vector transduction efficiency in mice in vivo and observed that at 4-weeks post-vector administration, liposomes had little effect on the transduction efficiency of rAAV2 or rAAV8 vectors, whereas cationic polymers significantly inhibited liver-targeted transgene expression by both serotype vectors. Taken together, these studies underscore the need to exercise caution in interpreting transgene expression data obtained following rAAV vector-mediated transduction, especially when additional transgene products are delivered by TAs.

305. Development of a qPCR Based Assay for Assessing Integrity of Oversized AAV Vectors

Heikki T. Turunen,1 Ru Xiao,1 Eva Plovie-Buys,1 Luk H. Vandenberghe.1 1 Gene Transfer Vector Core, Schepens Eye Research Institute, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, MA. In vivo gene transfer using AAV is emerging as a platform technology for the treatment of several forms of acquired and inherited disorders. One of the main limitations to it broad application is the genome size constrained imposed by the architecture of the viral particle. Efforts to map maximum genome capacity have ranged from 4.75kb up to 8.9kb. Data suggest that preparations exceeding wild type genome size are impacted in terms of genomic and structural integrity as is reflected in depressed yields and heterogeneity. While several biochemical and molecular methods are available for qualitatively assessing this issue, quantitative assays have to date not been developed. AAV genome packaging is initiated from either of the two identical ITR sequences flanking the genome. With normal sized genomes packaging proceeds in 3’ to 5’ direction until the opposite ITR is reached, whereas with oversized genomes packaging is terminated before the opposite ITR is reached, resulting in fragmented genomes lacking the distal ITR. As the ITRs are identical, a 1:1 ratio of (+) and (-) strand genomes is expected. However, we have detected a transgene sequence derived bias in genome packaging resulting in unequal ratios of (+) and (-) strands. Based on this observation we developed a TaqMan qPCR assay to study packaging of oversized genomes for rAAV vectors. As PCR is a template strand independent reaction, normal sized rAAV genomes give equal titers from any part of the genome regardless of the ratio of (+) and (-) genomes as templates. We performed TaqMan qPCR targeting promoter (5’) and polyA (3’) sequences for a normal sized (4.6 kbp) rAAV vector, and equal titers were detected, as expected. We then modified the construct to make it oversized (6.8 kbp) while retaining positions and sequences for the qPCR assay targets. Vectors were produced and titered identically to the normal sized vectors, and the 5’ qPCR gave 4 times higher signal than the 3’ qPCR. Similar results were obtained with AAV2, AAV5 and AAV8 capsids. Effects of the ITRs or non-packaged cisplasmid sequences to the strand-specific packaging were ruled out by confirming ITR integrity by sequencing, and by inverting the transgenic construct sequence between the ITRs and repeating the assay. Therefore, identical titers obtained from the two qPCR assays indicate that the whole genome has been packaged, whereas different titers indicate fragmented templates and thus oversized genomes. We are currently studying the effect of genome size to the ratio of (+) and (-) strands and developing an assay to study packaging of S117