Trans-scleral delivery of antiangiogenic proteins.

PURPOSE In this study, we investigated the penetration of various proteins into the mouse eye after a periocular injection of the protein or an adenoviral vector (Ad) expressing the protein. METHODS At several time points after the injection, the retina, retinal pigmented epithelium/choroid, and sclera were dissected and enzyme-linked immunosorbent assays were performed. RESULTS After a periocular injection of AdsFlt-1.10, AdTGFbeta.10, or AdPEDF.11, choroidal levels of pigment epithelium-derived factor (PEDF) and transforming growth factor-beta (TGF-beta) were not significantly different from scleral levels, and choroidal levels of sFlt-1 (soluble Flt-1 or soluble VEGF receptor 1) were only moderately reduced from scleral levels, indicating that each of these proteins penetrate the sclera well. In contrast, retinal levels of each of the three proteins were low compared to choroidal levels, suggesting poor penetration into the retina. Levels of PEDF in the choroid peaked 2 h after a periocular injection of PEDF protein and returned to baseline between 6 and 24 h, and peak levels in the retina were 8.6% of peak choroidal levels. Levels of green fluorescent protein, a protein unlikely to have any binding sites in mouse tissues, peaked in the choroid 2 h after the periocular injection and were undetectable by 4 h, while peak levels in the retina were 64.3% of peak choroidal levels. CONCLUSIONS These data suggest that size and binding characteristics of proteins are likely to influence their ability to penetrate the eye from the periocular space, but in general, proteins as large as 50-75 kDa penetrate well into the choroid, but not into the retina. Periocular injections are feasible for the treatment of choroidal neovascularization with proteins or vectors that express them, but additional investigations are needed before they can be considered for treatment of retinal diseases.