purpose. In the present study, the authors developed novel models to stimulate blood vessel formation (hemangiogenesis) versus lymphatic vessel formation (lymphangiogenesis) in the cornea.

methods. Micropellets loaded with high-dose (80 ng) or low-dose (12.5 ng) basic fibroblast growth factor (bFGF) were placed in BALB/c corneas. Angiogenic responses were analyzed by immunohistochemistry to quantify blood neovessels (BVs) and lymphatic neovessels (LVs) to 3 weeks after implantation. Areas covered by BV and LV were calculated and expressed as a percentage of the total corneal area (percentage BV and percentage LV). Hemangiogenesis (HA) and lymphangiogenesis (LA) were also assessed after antibody blockade of VEGFR-2 or VEGFR-3

results. Although high-dose bFGF stimulation induced a more potent angiogenic response, the relative LV (RLV= percentage LV/percentage BV× 100) was nearly identical with high-and low-doses of bFGF. Delayed LA responses induced 3 weeks after implantation of high-dose bFGF resulted in a lymphatic vessel-dominant phenotype. Interestingly, the blockade of VEGFR-2 significantly suppressed BV and LV. However, the blockade of VEGFR-3 inhibited only LV (P= 0.0002) without concurrent inhibition of BV (P= 0.79), thereby resulting in a blood vessel-dominant phenotype

conclusions. An HA-dominant corneal phenotype can be obtained in BALB/c mice 2 weeks after implantation of an 80-ng bFGF micropellet with VEGFR-3 blockade. Alternatively, an LA-dominant corneal phenotype can be obtained 3 weeks after implantation of an 80-ng bFGF micropellet without supplementary modulating agents. These models will be useful in evaluating the differential contribution of BV and LV to a variety of corneal abnormalities, including transplant rejection, wound healing and microbial keratitis.