Abstract
Purpose :
For the purpose of developing an optimal treatment for NMNAT1-associated disease, a severe and early onset inherited retinal degeneration (IRD), we are testing multiple variables related to AAV-based gene therapy. Here we report results from experiments in which the effects of different AAV serotypes, AAV doses, and components of the injection solution were evaluated in Nmnat1 mutant mice.
Methods :
Multiple AAV-NMNAT1 gene therapy vectors were developed that contain a human NMNAT1 cDNA and additional elements related to transgene expression. Vectors differ by serotype and/or promoter sequence. For treatment experiments, 0.75µL of solution containing the AAV-NMNAT1 vectors were injected into the sub-retinal space of one eye of two-week-old Nmnat1-V9M mice or littermate controls. The standard injection buffer contained PBS, 0.001% Pluronic F68 (surfactant), an additional 35mM NaCl, and 0.1% fluorescein. For example, NMNAT1 driven by the CAG promoter was packaged into AAV2/Anc80 and delivered at 1.0x108, 5.5x108, 1.0x109, or 3.4x109 genomic copies per µL (gc/µL). Separate experiments were performed using dilution buffer containing 0.001%, 0.01%, or 0.1% surfactant. Retinal morphology was assessed at different time points following injection using fundus imaging, OCT, immunolabeling, and standard light microscopy.
Results :
Some doses of AAV-NMNAT1 prevent or reduce retinal degeneration. For example, at four weeks post-injection, photoreceptor layer thickness in mutant retinas treated with 5.5x108 gc/µL and untreated wildtype retinas are nearly equivalent (100 µm vs. 110 µm, p=0.03). The photoreceptor layer of the treated mutant retinas is ~2x thicker than that of untreated fellow retinas (58 µm, p=0.001) and injected wildtype retinas are unaffected. However, higher doses (e.g., 3.4x109 gc/µL) of AAV-NMNAT1 cause persistent retinal detachment and/or accelerated degeneration. This damage is not associated with injection of AAV-EGFP or modified injection solution.
Conclusions :
While gene augmentation for NMNAT1-associated IRD has the potential to be therapeutic, high doses of the reagent appear to be toxic. Therefore, optimization of treatment parameters (e.g., by broad dose-response assessment) likely will be needed prior to clinical application. This finding may also be relevant to the development of other gene therapies for IRDs in addition to NMNAT1-associated disease.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.