Retinal microglia observed in human AMD specimens have amoeboid morphologies, suggesting their activated status. The initiating steps in this proposed model of cumulative subretinal microglia accretion have not been defined but may potentially be related to complement activation, drusen accumulation, photoreceptor injury, or oxidative stress in the outer retina. Further studies on the regulatory mechanisms underlying microglia distribution and migration in the retina, as well as their activation in the subretinal space, may be useful in this respect. Our results also indicated that retinal microglia co-culture increased in RPE cells the expression and secretion of VEGF and pro-angiogenic metalloproteinases, MMP1, MMP2, and MMP9. Co-culture RPE supernatants were also effective in increasing endothelial cell proliferation and migration as evaluated in 3 separate in vitro assays. These findings, together with an increase in pro-inflammatory and chemoattractive factors, suggest that microglial accumulation in the subretinal space may create a pro-angiogenic environment that increases the likelihood of the development of CNV. Metalloproteinases, thought to be Ginsenoside-F4 capable of comprising the natural barrier function of Bruch��s membrane to CNV growth, have been previously related to the formation of CNV in AMD and in animal models. The growth of CNV may be further Lomitapide Mesylate encouraged by the increased secretion of VEGF by RPE cells following co-culture that can directly stimulate endothelial cell proliferation and migration. In support of the relationship betwen RPE alterations and CNV formation, epidemiological studies have also linked the clinical presence of RPE changes occurring in patients with early and intermediate AMD with an elevated 5-year risk for the development of CNV. Our experiments with the in vivo transplantation of retinal microglia demonstrated that the presence of subretinal microglia exerts a strong pro-angiogenic influence in the formation and growth of CNV. The creation of a pro-angiogenic environment in the subretinal space, while mediated significantly by RPE alterations, may also be contributed towards by the subretinal microglia themselves. In our in vitro co-culture experiments, while mRNA and protein analysis in RPE cell lysates relate directly to RPE gene expression, protein analyses involving RPE co-culture supernatants may however also contain contributions secreted from retinal microglia cells. However, in our angiogenesis and microglia-migration functional assays, we found that the changes induced by RPE co-culture supernatants were significant when compared not only to unexposed RPE supernatants, but also when compared to supernatants of activated retinal microglia not exposed to RPE cells. As such, despite possible contributions by retinal microglia themselves, changes in RPE gene expression and protein secretion are likely play a prominent role in inducing the structural and functional changes of pathological significance in the subretinal space. While our in vivo model system permits direct cell-cell contact between retinal microglia and RPE cells, the in vitro system used here brings the two cell types in close proximity but however precludes direct cellular surface contact. The similarities in the nature of changes induced in both in vitro and in vivo systems suggest that many of these do not require direct microglial-RPE cell contact, although additional studies may be required to investigate the effects that result only from direct cellular contact. In addition to retinal microglia-to-RPE communication, other forms of intercellular interactions may also play a role in AMD pathogenesis.