VEGF Trap(R1R2) suppresses experimental corneal angiogenesis

Abstract

Purpose: To determine the effect of vascular endothelial growth factor (VEGF) TrapR1R2 on bFGF-induced experimental corneal neovascularization (NV).

Methods: Control pellets or pellets containing 80 ng bFGF were surgically implanted into wild-type C57BL/6 and VEGF-LacZ mouse corneas. The corneas were photographed, harvested, and the percentage of corneal NV was calculated. The harvested corneas were evaluated for VEGF expression. VEGF-LacZ mice received tail vein injections of an endothelial-specific lectin after pellet implantation to determine the temporal and spatial relationship between VEGF expression and corneal NV. Intraperitoneal injections of VEGF TrapR1R2 or a human IgG Fc domain control protein were administered, and bFGF pellet-induced corneal NV was evaluated.

Results: NV of the corneal stroma began on day 4 and was sustained through day 21 following bFGF pellet implantation. Progression of vascular endothelial cells correlated with increased VEGF-LacZ expression. Western blot analysis showed increased VEGF expression in the corneal NV zone. Following bFGF pellet implantation, the area of corneal NV in untreated controls was 1.05+/-0.12 mm2 and 1.53+/-0.27 mm2 at days 4 and 7, respectively. This was significantly greater than that of mice treated with VEGF Trap (0.24+/-0.11 mm2 and 0.35+/-0.16 mm2 at days 4 and 7, respectively; p<0.05).

Conclusions: Corneal keratocytes express VEGF after bFGF stimulation and bFGF-induced corneal NV is blocked by intraperitoneal VEGF TrapR1R2 administration. Systemic administration of VEGF TrapR1R2 may have potential therapeutic applications in the management of corneal NV.

Figures

Figure 1

Temporal and spatial relationship of VEGF expression and corneal vessel formation. Mouse corneas were implanted with a bFGF pellet and photographed by slit lamp on day 1 (D), day 4 (F), day 7 (H), day 10 (J), day 14 (L), and day 21 (N). A blank pellet was implanted as the control (A). bFGF pellet localization was shown in the transversal eye (B) and the intrastromal section (C). Sections of corneas were stained with anti-VEGF and anti-CD-31 antibodies on days 1 (E), 4 (G), 7 (I), 10 (K), 14 (M), and 21 (O). VEGF expression was noted in the corneal epithelium at day 1 (E). VEGF expression in the corneal keratocytes peaked on day 7 (I) and its expression decreased after 7 days. CD-31 localization lagged behind VEGF expression, which started on day 4 and continued until day 14. (* asterisk indicates the location of the bFGF pellet; arrows point to the limbus.) Areas of VEGF expression are stained red with anti-VEGF antibody; vascular endothelial cells are stained dark blue with anti-CD31 antibody.

Figure 2

VEGF expression correlated with vascular progression in the cornea. Vascular progression was induced by bFGF pellet implantation on days 1 (A), 4 (B), and 7 (C). Vascular endothelial cells were visualized by fluorescein-conjugated tomato lectin, and the progression of vessels was visualized by the VEGF-LacZ expression on days 1 (D), 4 (E), and 7 (F). LacZ expression (in red) was observed starting at day 4 and showed greater expression at day 7 (E–F). bFGF implanted corneas were divided into 6 segments as illustrated in (G) and were analyzed by Western blot analysis (H). Fifteen kDa bands (H, lanes 1 to 6) corresponding to VEGF expression in different segments (G, lanes 1 to 6) were observed. Recombinant VEGF was used as control (H, lane 7). The highest amount of VEGF was seen close to the pellet (H, lane 3) and in the segments adjacent to the pellet (H, lanes 2 and 4). There is also a lighter band noted at the limbal area closer to the pellet (H, lane 1). No VEGF expression was observed in segments away from the pellet (H, lanes 5 and 6). bFGF implanted corneas were coimmunostained with anti-VEGF antibody (I) and macrophage marker F4/80 antibody (J; merged image (K)).

Figure 3

VEGF-TrapR1R2 blocks bFGF-induced corneal NV. Mice were intraperitoneally injected with VEGF TrapR1R2 or hFc-protein before 80 ng bFGF pellet implantation. Corneal NV was photographed at day 4 (A, B, and C) and day 7 (D, E, and F). The distance of the pellet to the limbus was measured by 3 observers for each corneal images as described in the Materials and Methods section. There were no significant difference in the distance of pellets to limbus between 3 groups (control = 1.01mm±0.17mm; hFc = 1.21mm±0.22mm; VEGF Trap = 1.10mm±0.20mm; p=0.30). Enhanced corneal NV was documented in bFGF-implanted corneas with hFc-protein injection (B, E) and without peptide injection (A, D). bFGF-induced corneal NV was blocked by intraperitoneally injected VEGF TrapR1R2 (C, F). At day 4 after bFGF pellet implantation, the areas of corneal NV in these three groups (bFGF pellet only, or combined with human Fc injection or VEGF TrapR1R2 injection) were calculated and compared (G).