Apoptosis-inducing factor (AIF) is a novel mediator in apoptosis. AIF is a flavoprotein that is normally confined to the mitochondrial intermembrane space, yet translocates to the nucleus in several in vitro models of apoptosis. To investigate the role of AIF in the apoptotic process in vivo, we induced retinal detachment (RD) by subretinal injection of sodium hyaluronate, either in Brown Norway rats or in C3H mice. Apoptotic DNA fragmentation, as determined by terminal nick-end labeling, was most prominent 3 days after RD. The subcellular localization of AIF was examined by immunohistochemistry and immunoelectron microscopy. In normal photoreceptor cells, AIF was present in the mitochondrion-rich inner segment. However, AIF was found in the nucleus after RD. Photoreceptor apoptosis developed similarly in C3H control mice, and in mice bearing the gld or lpr mutations, indicating that cell death occurs independently from the CD95/CD95 ligand system. Both the mitochondrio-nuclear transition of AIF localization and the nuclear DNA fragmentation were inhibited by subretinal application of brain-derived neurotrophic factor. To our knowledge, this is the first description of AIF relocalization occurring in a clinically relevant, in vivo model of apoptosis.

The present results showed that apoptosis was a key phenomenon in the retinal dysfunction in RD and that this process was transmitted mainly by mitochondria-dependent pathways rather than Fas/Fas-L or downstream caspase dependent pathways.

A novel method of in vivo targeted gene transfer to intenthelial cells as early as day 1 and lasted until day 21. The tionally selected areas of the corneal endothelium was most intense gene expression was observed on days 1 and developed. Plasmid DNA with the lacZ gene coding for ␤-3 (5.21% on day 1 and 6.45% on day 3). The expression galactosidase was injected into the anterior chamber of of ␤-galactosidase on day 3 was most evident following adult Wistar rats, and eight pulses of electricity at intendelivery of 20 V electric pulses (0.09% at 5 V, 0.03% at sities ranging from 5 to 40 V/cm were delivered for 50 ms 10 V, 6.45% at 20 V). ␤-Galactosidase expression was limto the cornea with a specially designed electric probe in ited to the corneal endothelial cells in highly selected areas order to determine the effect of gene transfer on the corand no ␤-galactosidase expression was detected in any neal endothelial cells. Gene expression was visualized by other intra-or extraocular tissues. In addition, no cell damenzymatic color reaction using X-gal in enucleated eyes on age was apparent in the cornea and no inflammation was days 1, 3, 7, 14 and 21 after gene transfer. The treated detected in any other intraocular tissues. Thus, low-voltage eyes were then photographed and the X-gal-positive areas electric pulses successfully transferred the gene of interest were evaluated by an image analyzer. The ratios of the to highly selective areas of the corneal endothelium without areas (X-gal-positive area/area of entire corneal endoinducing any pathological changes. This targeted gene thelium × 100%) were then calculated to determine gene transfer method appears to have great potential for use in transfection efficiency. The expression of ␤-galactosidase gene therapy for ocular diseases. was clearly detected in the cytoplasm of the corneal endo-