In vivo confocal imaging of corneal neovascularization

Purpose. To investigate the cellular dynamics of vessel formation during corneal neovascularization in the living eye by confocal microscopy. Methods. Corneal neovascularization was initiated by placing a 7-0 silk suture through the corneal stroma 3 mm from the limbus at the 12 o'clock position in both eyes of 10 New Zealand white rabbits. The corneas were examined for vessel ingrowth at intervals from 1 to 15 days after suture placement using a tandem scanning confocal microscope with a 20X water immersion objective, as well as a slit-lamp biomicroscope. Changes in the limbal vessels were recorded on videotape for later analysis. As early vessel growth appeared to be associated with corneal nerves, the total number of sprouts and the number of sprouts along nerves were counted in confocal images, and the results analyzed for statistical significance. Vessel growth and the structural relationship between vascular buds and the deep stromal nerves were examined by light and transmission electron microscopy. Results. The early events of cell migration from the limbal microvessels were found to be associated with the deep stromal nerves; although this association was easily visualized by confocal microscopy, it could not be documented by slit-lamp biomicroscopy. By 18 h after suture placement, the limbal vessels were dilated and the first vascular buds appeared as short, pointed, or flat-topped protrusions from the deep limbal capillaries. By 96 h, the capillary buds had increased in density and had begun to form lumens. Movement of red blood cells was established between 72 and 80 h after the first signs of bud formation, at the same time that cells of immune origin were seen. Confocal microscopy revealed and transmission electron microscopy verified that new bud formation began with the formation of vascular tubes by endothelial migration along the deep stromal nerves. The total number of sprouts and the number of sprouts associated with stromal nerves were similar on days 1 and 2 but differed on days 3-7, suggesting an association between sprouts and nerves in the early stages of neovascularization. Conclusion. Using real-time white light confocal microscopy, we were able, for the first time, to observe the process of corneal neovascularization in the living eye, from the earliest stages within hours after initiation to 2 weeks. The deep stromal nerves appear to serve as a focus for the growth of new vessels, by attracting and supporting vessel growth and/or by providing a potential space for movement of the endothelial cells. Confocal microscopy may provide a new approach to achieving a better understanding of the mechanisms involved in corneal neovascularization.