Scholarly article on topic '241. Vector Design Tour De Force: Integrating Combinatorial and Rational Approaches To Derive Novel Adeno-Associated Virus (AAV) Variants'

241. Vector Design Tour De Force: Integrating Combinatorial and Rational Approaches To Derive Novel Adeno-Associated Virus (AAV) Variants Academic research paper on "Biological sciences"

Share paper
Academic journal
Mol Ther
OECD Field of science

Academic research paper on topic "241. Vector Design Tour De Force: Integrating Combinatorial and Rational Approaches To Derive Novel Adeno-Associated Virus (AAV) Variants"

to transduce motor neurons in adult spinal cord, which will greatly facilitate the evaluation of treatment strategies targeting this hard-to-reach population of cells in adult mice, which can be readily translated into clinical application. Currently, we are evaluating intracerebroventricular infusion of rAAV vectors using osmotic pumps, aiming to achieve enhanced global CNS transduction by delivering rAAV into CSF space.

239. Enteric Nervous System Transduction Following Intravascular Delivery of AAV9

S. E. Gombash Lampe,1 C. J. Cowley,1 F. L. Christofi,1 J. A.

Fitzgerald,1 K. D. Foust.1

'The Ohio State University, Columbus.

Gastrointestinal (GI) dysfunction is frequently reported in patients with neurologic and neuromuscular diseases. Symptoms related to GI dysmotility, such as constipation, gastroparesis and gastroesophageal reflux, impair the quality of life for patients with Parkinson's disease, multiple sclerosis, spinal muscular atrophy, spinal cord injury, autism spectrum disorders, and dementia. Peristalsis is highly regulated by the enteric nervous system, composed of the myenteric and submucosal nerve plexuses. Therefore, development of novel therapeutics that target enteric neurons may be of great benefit in treating GI symptoms. The purpose of the current study was to characterize transduction efficiency and transgene distribution in the myenteric plexus of the complete GI tract following intravenous injection of adeno-associated virus serotype 9 expressing green fluorescent protein (scAAV9-CB-GFP) in neonatal (P1) and juvenile (P21) mice. Neonatal mice were intravenously injected with scAAV9-CB-GFP (3 x 1011 vg) in the temporal face vein on P1 and juvenile mice received tail vein injections on P21 (2 x 1012 vg). Mice were euthanized at 60-90 days of age and the complete GI tract was dissected to expose the myenteric plexus. Immunohistochemical labeling of the myenteric plexus revealed GFP immunoreactivity in the esophagus, stomach, small intestine, cecum and colon in mice injected at each age. Neuronal transduction ranged from 21-57% depending on GI region analyzed. GFP colocalized with HuD positive myenteric neurons and S100 positive glial cells. Characterization of specific neuronal subtypes is ongoing and co-labeling studies thus far are consistent with multiple types of neurons in the ENS. Quantification of GFP immunoreactive neurons throughout the GI tract revealed increased transduction in P21 injected mice compared to P1 injected mice in most regions. This is in contrast to CNS data where neuronal transduction is more prominent in neonate injected animals. Segmental examinations of myenteric neurons in the colon showed no oral-aboral bias. These findings are significant because it shows AAV9 can be used as a tool to study intestinal function and for the development of GI therapeutics. Comparisons with other AAV serotypes are ongoing.

240. Efficient and Targeted Transduction of Nonhuman Primate Liver With Optimized AAV3B Vectors Through Systemic Delivery

Li Zhong,1,2 Shaoyong Li,1,3 Mengxin Li,1,3 Chen Ling,4 Qin Su,1 Ran He,1 Arun Srivastava,4 Guangping Gao.1,3 'Gene Therapy Center, UMMass Medical School, Worcester, MA; 2Pediatrics, UMass Medical School, Worcester, MA; Microbio & Physiol Systems, UMass Medical School, Worcester, MA; 4Pediatrics, University of Florida, Gainesville, FL.

Recent studies have demonstrated that rAAV3B vectors transduce human liver cancer cells as well as primary human hepatocytes efficiently in vitro. The combinations of modifications of specific surface-exposed serine (S) and threonine (T) residues on rAAV3 capsids further augment transduction efficiency in human liver cancer cells. Moreover these optimized rAAV3B vectors can achieve targeted delivery in a human liver cancer xenograft model after systemic

administration, suggesting that AAV3B and their derivative vectors are ideal for liver-directed gene therapy in humans. Here, we used the nonhuman primate model for liver-directed gene transfer to examine the safety and efficacy of these vectors which is critically important for their further clinical development in liver gene therapy in humans. To this end, we first investigated sero-epidemiology ofAAV3B in a rhesus monkey colony and found that the pre-existing neutralized antibodies (Nab) against AAV3B are relatively low in rhesus macaques (i.e. 52% of animals with no detectable Nab at a sensitivity of 1:5). To study liver tropism of optimized AAV3B vectors in NHP liver, we dosed three young (3-4 months) male rhesus macaques with S663V+T492V modified AAV3B (AAV3B.ST)-CB-EGFP vector at 1x1013 GC/kg intravenously and analysed EGFP expression and histopathology in the liver 7 days later by fluorescence microscopy. The results revealed that the optimized AAV3B.ST vector led to efficient and specific EGFP expression in the liver (36% hepatocytes in average as compared to 25% achieved by AAV7 in NHP liver) but not in other tissues, and without apparent hepatotoxicity. To quantitatively assess transduction efficiency of AAV3B and optimized AAV3B.ST vectors in NHPs, we i.v. injected three young male rhesus macaques with wild type (WT) and ST modified AAV3B-CB-rhCG vectors at 1x1013 GC/kg respectively and monitored serum rhCG levels for 3 months. Both WT and ST modified AAV3B vectors led to efficient and sustained rhCG expression in 104 - 105 rU/ml range up to 91 days, but ST modification enhanced rhCG expression by 2 (at late stage) to 5 (at early stage) folds as compared to WT AAV3B vector. This was confimed with qRT-PCR analysis of liver rhCG mRNA levels at the 91 day end point, showing again a 2-fold increased transcription by ST modification. The biodistribution analysis of the persisting vector genomes indicated a predominant liver targeting. Clinical chemistry and histopathology examinations of the study animals showed no apparent vector-related toxicity. In summary, the optimal AAV3B.ST vector can safely and specifically delivery transgene to primate liver and resulted in a further increased transduction efficiency as compared to WT AAV3B vectors after systemic administration. Our studies should be important and informative for clinical development of optimized AAV3B vectors for liver-directed gene therapy in humans.

* Equally contributed authors

# Co-corresponding authors.

241. Vector Design Tour De Force: Integrating Combinatorial and Rational Approaches To Derive Novel Adeno-Associated Virus (AAV) Variants

Damien Marsic,1 Lakshmanan Govindasamy,2 Seth Currlin,1 David Markusic,1 Yu-Shan Tseng,2 Roland W. Herzog,1 Mavis Agbandje-McKenna,2 Yuan Lu,3 Steve Ghivizzani,3 Sergei Zolotukhin.1 'Pediatrics, University of Florida, Gainesville, FL; 2Biochemistry and Molecular Biology, University of Florida, Gainesville, FL; 3Orthopaedics & Rehabilitation, University of Florida, Gainesville, FL.

AAV-derived vectors are promising tools for human gene therapy applications because of their absence of pathogenicity, episomal localization and stable transgene expression. However, significant limitations to their clinical use are their lack of specificity and their susceptibility to neutralization by human antibodies. Both of these limitations are determined by the nature of amino acid residues exposed at the surface ofthe capsid. Methodologies to improve existing AAV vectors for gene therapy include either rational approaches or directed evolution to derive capsid variants characterized by superior transduction efficiencies in targeted tissues. Although both methodologies have been successful in creating improved vectors, their utility is limited by our understanding of the AAV life cycle and by the technical boundaries of the protocols for directed evolution. The strategy adopted in this project integrates both approaches in one unified design. We have conducted a process of "virtual family

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

shuffling" in silico to derive a combinatorial capsid library whereby only variable regions on the surface of the structure are modified. An empirical step of pre-selection for 3D structural compatibility was introduced by first assembling individual sub-libraries in order to minimize the generation of dead-end variants. Next-Gen sequencing of the packaged viral DNA revealed capsid surface areas susceptible to directed evolution thus providing guidance for future designs. We demonstrated the utility of the library by deriving an AAV2-based vector characterized by at least 20-fold higher transduction efficiency in murine liver. This novel AAV liver-targeted capsid variant was evaluated by in vivo expression of luciferase and hFIX, and performed as well as the "golden standard" AAV8 in murine liver gene transfer. In another directed evolution experiment using the same capsid library, a novel AAV capsid variant targeting human osteosarcoma was derived, exhibiting high specificity and 10-fold higher transduction efficiency for tumor cells after systemic delivery in xenograft mouse model.

242. Genomic Analyses Define a Common Integration Site in Hepatocellular Carcinoma Caused By Therapeutic AAV Gene Delivery: A Mouse-Specific Phenomenon or New Safety Challenge for Human Gene Therapy? Randy J. Chandler,1 Adi A. Ashok,1 Gaurav K. Varshney,1 Matthew C. Lafave,1 Weiwei Wu,1 Victoria Hoffmann,1 Abdel G. Elkahloun,1 Shawn M. Burgess,1 Charles P. Venditti.1

'National Human Genome Institute, National Institutes of Health, Bethesda, MD.

Numerous toxicology studies have demonstrated the safety of AAV vectors. However, Donsante et al. (Science 2007) have reported an increased incidence of hepatic cellular carcinoma (HCC) after treatment of a murine model of mucopolysaccharidosis type VII with AAV2 in the neonatal period and implicated insertional mutagenesis of the vector at the Rian locus as a causative factor. We have reported pre-clinical AAV gene therapy studies for methylmalonic acidemia (MMA) and demonstrated the efficacy of neonatal systemic gene delivery. While evaluating long-term genotoxicity was not the intended focus of our studies, we observed that 75% (n=48) of AAV8-CBA-Mut treated mice developed HCC between 12 and 21 months, compared to a rate of 2% (n=41) noted in uninjected mice. In addition, 50% of the mice treated with AAV8-CBA-GFP (n=11) also developed HCCs, indicating that overexpression of the Mut transgene is not causative. We also performed an AAV dose escalation study that showed increasing the AAV8 dose from 7x1011-12 GC/kg (n=16) to approximately 1x1014 GC/kg (n=19) caused a increase in the occurrence of HCC from 12% to 84%(P<0.01). A high-throughput method (Varshney et al.Gen Res 2013) was used to capture and profile 2,350 unique and mapable integration events from 43 AAV-associated HCCs and 34 AAV-treated liver samples. In total, we observed a large number of integrations into liver-specific genes such as albumin (n=143) and alpha fetal protein (n=51). AAV integrations within a 5kb genomic region in the Rian locus (n=33), specifically either in or near miR34', had a significant correlation with HCC(P<0.01). Integration events were confirmed in 15 primary tumors by direct amplification and sequencing. Semi-quantitative PCR revealed that the HCCs had an AAV integration copy number between 0.25-1 per diploid genome; integrations in adjacent normal liver tissues were not detected. Expression analyses from HCCs and unaffected livers revealed an upregulation of non-coding RNAs from the Dlk'-Dio3 genomic region, which encompasses Rian. Even though it has been documented that mice treated with control and therapeutic AAV vectors developed HCC at significantly increased rates (Bell et al. Mol Ther 2006), the observations of Donsante et al. have remained controversial. Our data, the largest and most comprehensive set described to date, clearly establish that insertional mutagenesis at the Rian locus by AAV causes the development ofHCC

in a genotype and transgene independent-fashion. Furthermore, the tight clustering of integration events at the miR341 locus may have safety implications for clinical AAV gene delivery. While an ortholog for miR341 in humans has not been identified, deregulation of non-coding RNAs in DLK1-DIO3 region does occur in a subset of human HCCs. Determining why AAV gene delivery is rarely associated with increased tumorigenesis in mice and whether such toxicity translates across species will lead to improved safety and monitoring after AAV gene delivery in humans.

243. The Role of Heparan Sulfate Affinity in the Transduction of Photoreceptors By Intravitreally-Injected AAV2-Based Capsid Mutants

Sanford L. Boye,1 Antonette Bennett,2 Kim VanVliet,2 Dinculescu Astra,1 Miranda White,1 James Peterson,1 Mavis Agbandje-McKenna,2 Shannon E. Boye.1

Ophthalmology, University of Florida, Gainesville, FL; 2Biochemistry and Molecular Biology, University of Florida, Gainesville, FL.

Purpose: Developing AAV vectors capable of efficiently transducing retina from the vitreous would be a major step forward in translating gene therapy to the clinic. It is known that mutagenesis of surface exposed tyrosine residues prevents proteosomal degradation and increases nuclear transport of AAV thereby increasing its transduction efficiency. A variant with four Y-F mutations, AAV(quadY-F), transduces distal layers of the retina, including photoreceptors when delivered to the vitreous. The mechanism for this enhanced "penetration" is not fully understood. The primary cellular receptor for AAV2 is heparan sulfate (HS) and heparan sulfate proteoglycan (HSPG) is a major component of the inner limiting membrane. We hypothesized that the relative 'penetrating' ability of capsid mutants relies on their respective HS affinities.

Methods: AAV2- based vectors containing combinations of Y-F and/or T-V mutations were analyzed by chromatography on heparin agarose columns to determine respective affinities. A new set of capsid variants was created using a triple Y-F/single T-V variant (AAV2 MAX) as template. Structure informed mutations of known HS binding residues on the AAV2 capsid were made to generate vectors with a range of HS affinities or no affinity whatsoever (AAV2 MAX delta HS). New variants are currently being evaluated for transduction efficiencies in vitro and in vivo via subretinal and intravitreal injections in mice.

Results: AAV2 variants that primarily transduce inner retina from the vitreous (AAV2wt and AAV2tripleY-F) have strong affinity for HS, while those that transduce photoreceptors have moderate affinity (AAV2quadY-F and AAV2quadY-F+T-V). The variant with no HS affinity (AAV2 MAX delta HS), does not transduce cells in vitro or retina when delivered by intravitreal injection. However when delivered subretinally, this variant is highly efficient at transducing photoreceptors.

Conclusion: HS affinity of AAV2 based vectors is a key factor in their ability to transduce retina from the vitreous. Those with moderate HS affinities are capable of transducing photoreceptors after intravitreal delivery. Interestingly, highly efficient photoreceptor transduction occurred in the absence of HS binding. Another ligand-receptor pair is responsible for attachment of AAV2MAX delta HS to the surface of photoreceptors.

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