The study of ocular gene therapy has focused on preclinical work in a gene therapy trial of gyrate atrophy (GA), optimization of nonviral approaches to gene delivery in retinal pigment epithelial cells (RPE), and investigation of possible ocular toxicities of the use of E1-deficient adenoviral vector. Using an E1-deficient adenoviral vector, which encodes ornithine aminotransferase (OAT) (hCMV.Ad5.OAT) further studies have been performed on the replacement of this enzyme in isolated keratinocytes obtained from patients with GA. Although the rate of incorporation of 14C-ornithine into protein into these cells before transduction with hCMV.Ad5.OAT is essentially zero, following transduction the rate of incorporation exceeds that rate seen in normal keratinoctyes. This indicates a single, correctable defect in the ornithine pathway in GA cells. The rate of disappearance of 14C-ornithine from the media as measured by the rate of accumulation of 14C-ornithine-metabolites in GA keratinocytes overexpressing OAT is considerably higher than that seen in normal keratinocytes. This indicates that a """"""""metabolic sink"""""""" may be possible where a small patch of transduced keratinocytes in GA patients may be able to effectively reduce the elevated serum ornithine concentration seen in these patients. Human RPE cells in culture have been shown by our group to be readily transduced by adenovirus vectors. However, following transduction, these cells exhibited toxicity, which appeared to be vector related. Southern blot and dot/blot gels revealed a time-dependent replication of viral DNA in the absence of complementing E1 sequences. There was also a time-dependent increase in the amount of intracellular viral hexon protein as detected by immunohistochemistry. This phenomenon was present in vitro in both undifferentiated and differentiated RPE cells. Although this toxicity may limit the usefulness of first generation adenovirus vectors, its identification allows for careful screening of newer second and third generation adenovirus vectors. Because of the cellular toxicity observed in human RPE cells following adenovirus vector transduction, nonviral approaches to gene transfer into these cells have been explored. Standard lipofection and calcium transfection techniques result in low efficiency of gene transfer in human RPE cells. However, when an E1-deleted adenovirus vector without transgene is added to a lipofection protocol, the efficiency increases more than 10 fold. Because of the above-mentioned replication of this vector, unacceptable toxicity was observed. However, when transferrin or ultraviolet-attenuated adenovirus is added to the lipofection protocol, efficiency is increased without a concomitant increase in cellular toxicity. Studies are under way, to evaluate these various approaches to gene transfer in vivo in various animal models. Efficient expression of marker genes (b-galactosidase and green fluorescent protein) in retinal pigment epithelial cells in vivo following subretinal injection of adenovirus vectors encoding these genes has been performed in various strains of both mice and rats. Construction and generation of adenovirus vectors encoding vascular endothelial growth factor, acidic fibroblast growth factor, and a constitutively active form of transforming growth factor beta-1 and viral-interleukin-10 have been completed. Studies are under way to evaluate the effects of these proteins in the subretinal space of both normal rodents and those with ocular disease.
