306. Characterization of a Novel Adeno-Associated Viral Vector with Preferential Oligodendrocyte Tropism

AAV Vectors II infect xenotransplanted mice and are then replicated with either Adenovirus-19b or 36. Two in vitro screens were performed in either primary human muscle stem cells isolated and purified from patient biopsies, or differentiated human myotubes. Completed screens were carried out for six rounds of replicating selection with Adenovirus-5 and the five most highly selected variants from each screen were sequenced and vectorized into EF1a-driven GFP-luciferase dual expression vectors. We are validating these 10 highly selected variants with comparisons to parental serotypes with some known muscle tropism (AAVs 1,6,8,9), as well as establishing neutralizing antibody levels against pooled human immunoglobulins. Vectors are being tested in humanized muscle mice and various human muscle cell lines (both healthy and diseased) by four methods of administration (intra-muscular, intra-arterial, intra-venous and intra-peritoneal) at a variety of doses for robust transgene expression. Best candidates will be used to package broadly neutralizing antibodies against HIV for further pre-clinical validation as a vaccine delivery agent. More generally, we hope to provide the muscle gene therapy community with a panel of optimized capsids with enhanced tropism for numerous muscle cell types with broad potential applications.

304. Deriving Useful Data from Next-Gen Sequencing of AAV Capsid Libraries Damien Marsic,1 Sergei Zolotukhin.1 1 Pediatrics, University of Florida, Gainesville, FL.

The relatively short read lengths produced by the major Next-Gen sequencing platforms (up to 300 nt for Illumina, up to 200 nt for Ion Proton) are poorly suited for analyzing large combinatorial libraries of longer nucleotide sequences, such as those involving AAV capsid genes. Despite its much lower throughput and accuracy, the PacBio single-molecule real-time technology is currently the only option for long templates, with average read lengths of 10 to 15 kb. Using the Circular Consensus Sequencing (CCS) mode, in which template DNA fragments are circularized, allows a significant increase in accuracy due to the fact that each template is being sequenced multiple times. To interpret PacBio CCS data, we have previously reported a first version of the CapLib code, which was developed to identify variable regions in AAV combinatorial capsid libraries. DNA fragments, derived from purified DNA-containing AAV particles, 869 bp in length and including 27 variable nucleotide positions, were sequenced in CCS mode using the P6-C4 chemistry. A total of 26,897 reads were obtained, with a mean read length of 814 nt, a mean read quality of 0.9956 and a mean number of passes of 21.34. Only 5,456 reads had the correct size of 869 nt, and of these, only 1,638 had a sequence that matched the reference sequence, indicating that only 6% of reads were potentially error-free and that the vast majority had multiple insertions and deletions. In order to extract more useful information from the sequencing data, a new version of the CapLib software was developed. It is designed to correct sequencing reads in silico by assuming that constant nucleotide positions are wild-type and focusing on the detection of the variable positions. The premise was validated by Sanger sequencing of multiple clones, confirming that mutations were present only in the intended positions. Depending on the parameter values used, up to 14,000 reads could be recovered by CapLib 2. In addition to recovering PacBio CCS reads, CapLib 2 can also assemble Sanger sequencing data, translate recovered reads into protein sequences and perform detailed analyses of the dataset. It can also analyze clones resulting from directed evolution experiments and compare them with the original library.

305. A Screening Strategy for Selecting Recombinant Adeno-Associated Viral Vectors That Selectively Transduce Viral Infected Cells

Gustavo de Alencastro,1 Nicole K. Paulk,1 Katja Pekrun,1 Mark A. Kay.1 1 Department of Genetics and Pediatrics, Stanford University, Palo Alto, CA.

Recombinant adeno-associated viral (rAAV) vectors constitute one of the most promising tools for gene transfer. Desired properties include the lack of pathogenicity in humans, broad tissue tropism, ability to transduce both dividing and non-dividing cells, and the potential to establish long-term and stable transgene expression in quiescent tissues. Additionally, the majority of transduction events are episomal, making the risk of insertional mutagenesis lower than integrating vectors. Preclinical studies have successfully utilized rAAV vectors to combat viral infections after delivering short-hairpin RNA (shRNA) expression cassettes for knocking down viral coding genes in infected cells. A viral infection can also be reduced using shRNAs against host cellular genes, which are required for the life cycle of the pathogen. Although combining both the targeting of viral and cellular genes has the potential to increase the success in combating viral infections, promiscuous targeting of host cellular genes in the majority of cells that remain uninfected in a tissue may be detrimental. Therefore, we have designed a strategy to select for rAAV capsids that will selectively transduce infected over uninfected cells. The selection is based on the fact that most viral infections cause phenotypic changes in a cell that likely affect at least one of the many parameters involved in rAAV transduction. For proof-ofconcept, we are selecting for rAAV capsids that selectively transduce Hepatitis B virus (HBV)-infected hepatoma cells, in part because it was recently reported that HBV infection increases AAV2 and AAV8 transduction. We are currently characterizing a new highly diverse rAAV capsid library generated by shuffling of 18 parental capsids placed into a wild-type AAV genome. This library will be co-infected with wild-type Adenovirus-5 (helper virus) allowing the rAAV genomes to replicate in HBV-infected human cells. Isolation of replicating rAAV over successive passages in cultured cells will allow for stringent selection as all steps in AAV transduction, including cell entry, intracellular transport, nuclear entry and capsid uncoating, must occur in the selection scheme. After selection is completed, we will sequence and vectorize the most abundant capsids. These will be individually tested for their relative transduction of hepatoma cells that are identical except for their HBV infection status. Identifying such a rAAV capsid specific for HBV-infected cells will provide a reagent or therapeutic that can deliver a transgene that will allow one to selectively modify or potentially kill infected but not the uninfected cells contained within the liver. Moreover, we believe this strategy has the potential to be broadly applied for the selection of rAAV capsids specific for cells infected with other viruses or any disease state that results in phenotypic changes to a cell.

306. Characterization of a Novel AdenoAssociated Viral Vector with Preferential Oligodendrocyte Tropism

Sara K. Powell,1 Nadia Khan,1 Christina Parker,1 R. Jude Samulski,1 Glenn Matsushima,2 Steven J. Gray,1 Thomas J. McCown.1 1 Gene Therapy Center, University of North Carolina, Chapel Hill, NC; 2Microbiology and Immunology, University of North Carolina, Chapel Hill, NC. An AAV vector with oligodendrocyte-centric tropism would provide a valuable tool for research or for gene therapy in the treatment of demyelinating neurological disorders, such as ALS and

Molecular Therapy Volume 23, Supplement 1, May 2015 Copyright © The American Society of Gene & Cell Therapy

S123

AAV Vectors II multiple sclerosis. Using AAV capsid DNA shuffling we created a library of novel AAV vectors. The primary recovered AAV capsid clone (Olig001) exhibited a novel >95% tropism for rat, striatal oligodendrocytes when a chicken beta actin (CBH) promoter drove GFP expression. Furthermore, when Olig001 is administered intravenously it is highly detargeted from peripheral tissues such as the liver. Surprisingly, Olig001 shares 99.3% homology (7 amino acid differences) with AAV8 within the VP3 coding region (aa204738 using VP1 numbering); however, AAV8-CBH-GFP vectors are predominantly neurotropic. While AAV8 has some minimal tropism for oligodendrocytes, our in vitro studies using primary mixed glia cultures showed that Olig001 binds to oligodendrocytes with much higher affinity than AAV8, thus Olig001 exhibits a true gain-offunction. We sought to identify which amino acid(s) of Olig001 contribute to the oligodendrocyte-preferring tropism. Mutation of some or all of the four C-terminal amino acids of Olig001 (downstream of aa343 using VP1 numbering) back to AAV8 still produced an oligodendrocyte-preferring capsid. The N-terminal, mostly VP1/ VP2-specific, portion of the capsid is a chimeric mixture of AAV1, 2, 6, and 9, and our results suggest that this region of the capsid is conferring the oligodendrocyte preferring phenotype. In summary, we report the generation of a novel AAV capsid with a tropism and biodistribution profile distinct from any naturally-occuring AAV capsids. Moreover, our results suggest that the N-terminus of the capsid specific to VP1/VP2 is the major contributor to this unique oligodendrocyte tropism. Work supported by Michel J. Fox Foundation and Target ALS grants to TJM.

307. Rational Development of 12 Different AAV Serotypes as Scaffolds for Peptide Display

Kathleen Boerner,1,2 Eike Kienle,1 Anna Sacher,3 Lin-Ya Huang,4 Marina Bechtle,1 Ellen Wiedtke,1 Philipp Bayer,1 Hans-Georg Kräusslich,1,2 Mavis Agbandje-McKenna,4 Dirk Grimm.1 1 Cluster of Excellence CellNetworks, Dept. of Infectious Diseases/ Virology, Heidelberg University Hospital, Heidelberg, Germany; 2 German Center for Infection Research (DZIF), Heidelberg, Germany; 3Tumorvirus-Specific Vaccination Strategies, German Cancer Research Center, Heidelberg, Germany; 4Dept. of Biochemistry and Molecular Biology, University of Florida, Gainesville. Recombinant AAV vectors are some of the most attractive vehicles for therapeutic gene delivery available today, for many reasons. One is that AAV capsids are highly amenable to genetic engineering, permitting the creation and identification of vectors with novel properties including target cell specificities. One potent strategy is to insert a randomized peptide library into an exposed AAV2 capsid region, followed by iterative selection on desired cells to enrich single peptides binding to cell-surface receptors and hence mediating AAV transduction. However, several problems hamper the use of AAV2 as scaffold for peptide display: (i) neutralizing anti-AAV2 antibodies that are prevalent in the human population may interfere with transduction of AAV2 peptide mutants, and (ii) if not fully ablated, the AAV2 tropism for its genuine receptors will dampen retargeting efficiencies. Here, we aimed to overcome both problems by developing 11 other AAV serotypes as templates for peptide display, AAV1, 3-9, rh.10, po.1 and 12. We first engineered their capsid genes to contain unique restriction sites and then exploited these for insertion of 18 different oligonucleotides encoding peptides which were pre-selected in AAV2 on various cells. The resulting >200 AAV variants were subsequently produced as YFP-expressing vectors and screened in over 80 human or murine cells, including primary hepatocytes or keratinocytes, as well as T-cells and stem/iPS cells. Strikingly, many capsid-peptide combinations vastly outperformed the AAV2 counterparts, especially S124

those based on AAV serotypes 1, 7-9 and rh.10, and those carrying 7mer peptides with an NxxRxxx motif (x = any amino acid). In addition, we found that disruption of residue R585 in AAV2 boosted the potency of several peptide display mutants by up to two orders of magnitude, whereas the same mutation had no effect in the chimeric capsid AAV-DJ, despite a common region surrounding R585. Another puzzling notion was that some serotypes were inactivated by peptide display, tempting us to model all 12 capsids after insertion of a prototype peptide. This revealed that in contrast to our prediction from linear sequence alignments, the exact position of the displayed peptide differed by up to four residues between the 12 capsids. Indeed, once adjusted accordingly, AAV3 and AAV6 were also markedly enhanced upon peptide insertion. Collectively, our data imply an enormous potential of alternative AAV serotypes as scaffolds for peptide display and provide a set of vital guidelines for their further development as gene therapy vectors. To aid in this process, we also present a protocol for AAV spotting and drying in 96/384-well plates which permits their long-term storage and shipping, and thus facilitates their evaluation in other labs.

308. Inclusion of Heterologous ITRs in Dual AAV Vectors for Retinal Gene Therapy

Elisabetta Toriello,1 Pasqualina Colella,1 Andrea Sommella,1 Ivana Trapani,1 Andrea Maddalena,1 Alberto Auricchio.1,2 1 Telethon Institute of Genetics and Medicine, Pozzuoli, Italy; 2 Translational Medicine, Federico II University, Naples, Italy.

AAV cargo capacity limited to about 5 kb prevents their applications from treatment of inherited blinding conditions which require the transfer of genes with larger coding sequences. However, dual AAV vectors, each containing one of the two halves of a large gene expression cassette, are emerging as promising tools to overcome this limitation. However, dual AAV show lower levels of transgene expression than a single AAV vector and are associated with production of proteins shorter than expected from either the 5’- or 3’half AAV. These limitations can be in part overcome by improving the productive tail-to-head AAV genome concatemerization by including heterologous inverted terminal repeats (ITRs) at both ends of each dual AAV vector genome. We have generated dual AAV hybrid vectors with heterologous ITRs from AAV2 (ITR2) and AAV5 (ITR5) which are the most divergent among the various AAV serotypes. However, we found a significant reduction in the yields of AAV vectors with heterologous ITR2 and 5 when compared to those of vectors with homologous ITR2. In addition, the genome titer performed on ITR2 sequences results lower than that on a different part of the genome suggesting that ITRs might not be intact in AAV vectors with heterologous ITRs. Finally, the amount of full length protein relative to shorter was similar between vector with homologous and heterologous ITRs . We are currently testing dual AAV hybrid vectors with ITRs derived from other serotypes divergent from AAV2 which could overcome some of the limitations we have found using ITR5.

309. Optimization of Dual AAV Vectors for Gene Therapy of Inherited Retinal Diseases

Ivana Trapani,1 Sonia de Simone,1 Carolina Iodice,1 Pasqualina Colella,1 Settimio Rossi,2 Alberto Auricchio.1,3 1 Telethon Institute of Genetics and Medicine (TIGEM), Pozzuoli (NA), Italy; 2Eye Clinic, Multidisciplinary Department of Medical, Surgical and Dental Sciences, Second University of Naples, Naples, Italy; 3Medical Genetics, Department of Translational Medicine, Federico II University, Naples, Italy. Retinal gene therapy with adeno-associated viral (AAV) vectors is safe and effective in humans. However, AAV limited cargo capacity Molecular Therapy Volume 23, Supplement 1, May 2015 Copyright © The American Society of Gene & Cell Therapy