Intraocular route of AAV2 vector administration defines humoral immune response and therapeutic potential

PURPOSE Safety and efficiency are critical for successful gene therapy. Adeno-associated viral (AAV) vectors are commonly used for gene transfer in both human and animal studies. However, administration of AAV vectors can lead to development of neutralizing antibodies against the vector capsid, thus decreasing the efficiency of therapeutic gene transfer and preventing effective vector readministration. We investigated immune responses to different routes of ocular administration and readministration of AAV vectors, and the effect of previous exposure of AAV vector in one eye on the transduction efficacy of subsequent intraocular AAV-mediated gene delivery to the partner eye. METHODS We tested two vector systems. One contained a cDNA encoding a secreted pigment epithelial derived factor (PEDF) cDNA under the control of a Cytomegalovirus (CMV) enhancer and chicken beta-actin promoter (CBA; AAV2-CBA-PEDF) and was tested in a murine model of laser-induced choroidal neovascularization (CNV). The other vector contained a cDNA encoding the intracellular reporter green fluorescent protein (GFP) under the control of the same promoter (AAV2-CBA-GFP). Animals were divided into groups and received sequential injections at different combinations of either intravitreal or subretinal routes. CNV was evaluated by fluorescein angiographic choroidal flat-mount image analysis. The expression of GFP was analyzed in retinal sections by direct fluorescence imaging. Antibodies against AAV2 capsid and transgenes were analyzed by ELISA using serum samples collected before injection and different time points after the injection. Neutralizing antibodies were characterized by in vitro assays. RESULTS Various ocular compartments responded to AAV administration differently. Intravitreal administration of AAV vectors, which resulted in transduction of inner retina (primarily retinal ganglion cells), generated a humoral immune response against AAV capsid that blocked vector expression upon readministration via the same route into the partner eye. In contrast, it had no effect on vector readministered into the subretinal space of the partner eye. Additionally, subretinal administration of vector did not trigger any humoral immune response against AAV capsid, and had no effect on subsequent administration of vector either intravitreally or subretinally into the partner eye. CONCLUSIONS These findings have important clinical implications for the design of AAV-mediated ocular gene transfer for retinal diseases, particularly if both eyes require sequential treatment.