Seng Cheng

249. Preliminary Results of a Phase 1, Open-Label, Safety and Tolerability Study of a Single Intravitreal Injection of AAV2-sFLT01 in Patients with Neovascular Age-Related Macular Degeneration

Clinical Trial: www.clinicaltrials.gov NCT01024998Introduction: Vascular endothelial growth factor (VEGF) plays a critical role in the development of the neovascular AMD (wet AMD), the most common cause of new onset blindness in elderly adults. While intravitreal (IVT) administration of anti-VEGF agents are effective for improving vision in wet AMD patients, the need for repeated monthly IVT injections imposes a burden on patients and ophthalmologists which may limit the realization of the full benefits of VEGF inhibition. AAV2-sFLT01 is a replication defective, recombinant adeno-associated virus (serotype 2) vector (AAV2) that encodes sFLT01, a novel soluble chimeric protein composed of domain 2 of VEGF receptor 1 and the Fc region of human Immunoglobulin G type 1 (IgG1). We investigated the safety and tolerability of a single IVT administration of AAV2-sFLT01 in wet AMD patients. Methods: The study was a Phase 1, open-label,, dose-escalating, safety and tolerability study conducted in 4 sites in the USA. 5 cohorts of patients with BCVA of ≤20/100 in the study eye received a single IVT injection of one of four AAV2-sFLT01 doses (2×108, 2×109, 6×109, or 2×1010 vg). Patients in cohorts 1-4 had subfoveal disciform scarring, while cohort 5 patients had responsiveness to an anti-VEGF therapy within 12 months prior to enrollment. All patients were monitored for 52 weeks following the injection in the core study, and encouraged to participate in the extended follow-up program for up to 4 years. The primary safety endpoint was the occurrence of adverse events. Biological activity was evaluated by changes from baseline in sub/intraretinal fluid by optical coherence tomography (OCT), best corrected visual acuity (BCVA) and sFLT01 protein level detected in the aqueous fluid. Results: 19 patients with end-stage/advanced wet AMD (53% males, median age 77.0 years, median time since diagnosis of AMD of 5.4 years and median basal BCVA of 22.0 letters) were treated. All patients completed the 52-week study with 17 patients consenting to the 4 year extended follow-up study. Overall, the treatment was considered safe and tolerable with no dose-limiting toxicity observed in any dose level. 11/19 patients were expected to show an anti-VEGF response clinically. Of them, transient (2 patients) and sustained (4 patients) reduction in retinal thickness suggested a biological effect of the treatment. Aqueous sFLT01 protein expression was identified in 5 patients receiving 2×1010 vg. Post-treatment changes in BCVA did not correlate with other parameters of biological activity. Conclusion: Despite advanced disease state, a sustained reduction in retinal fluid in four patients provides evidence of biological activity supporting the concept that a sustained therapeutic anti-VEGF level can be achieved following a single intravitreal administration of an ocular gene therapy.

329. Using CRISPR/Cas9 as a Therapeutic Approach for Leber Congenital Amaurosis 10 (LCA10)

Introduction: Leber congenital amaurosis (LCA) is the most severe form of inherited retinal diseases with early onset of symptoms in the first year of life. The most frequent genetic cause of LCA, accounting for approximately 15% of all LCA cases in western countries, is a deep-intronic mutation c. 2991+1655A>G located in the intron 26 of human CEP290 gene. LCA caused by CEP290 mutation is known as LCA10. The intronic mutation of CEP290 generates a cryptic splice donor site, resulting in the inclusion of a pseudoexon that leads to a premature stop codon and a truncated protein. The size of human CEP290 cDNA (~7.4 kb) exceeds the cargo size (~4.8 kb) of recombinant adeno-associated viral vectors (rAAVs), which makes this gene challenging for gene replacement therapy. In this regard, we tested if we can use the new genome editing technology CRISPR/Cas9 as an alternative strategy for LCA10 by removing the intronic mutation of CEP290 and preventing the cryptic splicing.Methods: CRISPR/SpCas9 was employed to introduce the intronic splice mutation c. 2991+1655 A>G into HEK 293FT cells to create a cellular model for LCA10/CEP290. In this cellular model, we used guide RNA pairs coupled with SpCas9 to delete the intronic region flanking the intronic mutation.Results: The deep-intronic mutation c. 2991+1655 A>G was successfully introduced into 293FT cells through the homology directed repair (HDR) pathway of CRISPR/spCas9. CEP290 expression levels were markedly reduced in the mutant cell line. Using this cellular model, we identified three sgRNA pairs that could efficiently (>50%) delete the intronic mutation, significantly rescue wild-type CEP290 expression levels and reduce mutant CEP290 expression.Conclusions: Our results demonstrate that the paired sgRNAs coupled with SpCas9 are capable of removing the c. 2991+1655A>G mutation and are highly efficient in preventing the splicing of the mutant cryptic exon and restoring wild-type CEP290 expression.

301. AAV Capsid Engineering to Improve Transduction in Retina and Brain

Gene therapy vectors based on adeno-associated virus (AAV) are currently in clinical studies for numerous disease indications including Lebers congenital amaurosis, age-related macular degeneration, hemophilia, muscular dystrophy and Parkinsons disease. AAV vectors hold considerable promise as therapeutic agents; however there is potential to further improve the efficiency of AAV gene delivery and efficacy by making modifications to the AAV capsid. The AAV capsid can be engineered to incorporate mutations that alter its transduction activity, tropism, biodistribution and immunogenicity. We have constructed variant AAV vectors harboring a variety of capsid modifications including those that negate receptor binding and have tested these vectors in several tissues including the eye and brain. One variant, AAV2HBKO, is an AAV2 based vector containing mutations of critical amino acids known to be required for binding to its receptor, heparin sulfate proteoglycan. Interestingly, an AAV2HBKO vector delivering a secreted transgene, sFLT02, unexpectedly resulted in a 2-log increase in transduction compared to parental AAV2 when delivered subretinally to the mouse eye. Subretinal delivery of an AAV2HBKO vector expressing EGFP demonstrated that these capsid modifications resulted in an increase in photoreceptor transduction compared to the unmodified AAV2 vector. In contrast, the AAV2HBKO vector demonstrated a lack of transduction activity following intravitreal delivery to the mouse eye. In addition, we evaluated the transduction and tropism of AAV2HBKO in the mouse brain. In a head to head comparison with AAV2, the AAV2HBKO vector facilitated widespread striatal and cortical expression following an intrastriatal injection while AAV2-mediated expression was restricted to the site of injection. Similar to AAV2, the tropism of AAV2HBKO was primarily neuronal with little to no transduction of astrocytes or microglia. Biodistribution data suggests that this vector, when delivered systemically in the mouse, has significantly reduced liver transduction but a higher propensity to be delivered to skeletal muscle and heart compared to the wild-type AAV2 vector. We will present data evaluating the transduction activity, tropism and biodistribution of the AAV2HBKO variant. These studies illustrate the potential for improving the efficiency of AAV gene transfer via targeted capsid engineering.

556. Analytical Ultracentrifugation as an Approach to Characterize Recombinant AAV Vectors

Appropriately engineered recombinant adeno-associated viral (rAAV) vectors represent a novel class of biopharmaceutical drugs. The production of clinical-grade rAAV vectors for gene therapy would benefit from analytical methods that are able to monitor drug product quality with regard to homogeneity, purity and manufacturing consistency. Here, we demonstrate the novel application of analytical ultracentrifugation (AUC) to characterize the homogeneity of preparations of rAAV vectors. We show that a single sedimentation velocity run of rAAV vectors detected and quantified a number of different viral species, such as vectors harboring an intact genome, lacking a vector genome (empty particles) and containing fragmented or incomplete vector genomes. This information is obtained by direct boundary modeling of the AUC data generated from refractometric or UV detection systems using the computer program SEDFIT. Using AUC, we show that multiple parameters contributed to vector quality, including the AAV genome form (i.e., self-complementary vs single-stranded), vector genome size, and the production and purification methods. Hence, AUC is a critical tool for identifying optimal production and purification processes and for monitoring the physical attributes of rAAV vectors to ensure their quality.

507. Down-Regulation of the Epithelial Sodium Channel ENaC through Antisense Oligonucleotides

Sodium absorption across the airway epithelium is increased in cystic fibrosis (CF) patients and may contribute to abnormal water transport and accummulation of sticky sputum, a hallmark of the CF lung disease. Down-regulation of epithelial sodium channels (ENaC) may attenuate the CF lung disease. One approach to decrease ENaC function is the specific down-regulation of its mRNA. Single-stranded DNA antisense oligonucleotides (ssDNA) bind to mRNA sequence-specifically and trigger RNase-H mediated degradation of the ssDNA/mRNA hybrid. However, careful selection of the ssDNA sequence is required to obtain optimal mRNA reduction. Here, we compared five murine ENaC antisense DNAs (20 nucleotides each) for their ability to decrease ENaC mRNA in M1 cells, a murine kidney cell line that expresses ENaC abundantly. It has previously been demonstrated that the alpha subunit of ENaC is the most crucial for channel function and all five ssDNAs were therefore directed against this subunit. First, fluorescently labelled ssDNA was complexed with Lipofectamine 2000 and transfection efficiency determined. Interestingly, the ssDNA rapidly (after 30 min) accumulated in the nucleus of transfected cells and after 6 hrs >90% cells were transfected. The nuclear localisation of the antisense ODNs was consistent with their proposed site of action. Forty-eight hrs after transfection cells were harvested and mRNA was prepared for quantitative RT-PCR using Taqman. The relative degree of aENaC mRNA reduction varied between antisense molecules and ranged from 5 to 60% (n=6, p<0.05). The most efficient ssDNA was used for in vivo transfection. Genzyme lipid 67 (GL67) has been widely used for gene transfer in vivo, but not for antisense applications. We first assessed if GL67 complexed antisense ssDNA efficiently. Gel retardation studies showed that with a 6:8 molar ratio of cationic lipid:nucleotide almost all ssDNA was complexed (n=3). Light-scatter analysis showed that complexes had an average size of 390±24 nm, but varied in size (n=3). Nasal potential difference assays (PD) have been widely used to measure ion transport across murine nasal epithelium. A decrease in ENaC function would result in a drop in baseline PD, because sodium transport is the main component of basal ion transport in the nasal mucosa. Here, we complexed aENaC antisense or control ssDNA to GL67 (80 mg/mouse) and perfused these complexes into the mouse nose over 75 min via a thin plastic catheter. Twenty-four, 48 and 96 hrs after administration the PD was measured in individual cohorts of animals (n=10/group). After the PD measurements all animals were culled and Taqman RT-PCR carried out on nasal mRNA. We did not detect a drop in aENaC mRNA or changes in PD at any time-point. A critical factor for antisense function in the mouse nose is transfection efficiency. In preliminary experiments (n=3) using fluorescently-labelled ODNs we noted that transfection efficiency varied drastically from animal to animal, and will therefore require further optimisation.