Inhibition of mouse alkali burn induced-corneal neovascularization by recombinant adenovirus encoding human vasohibin-1

Abstract

Purpose: To evaluate the activity of recombinant adenovirus encoding human vasohibin-1 (Ad-Vasohibin-1) on mouse corneal neovasularization induced by alkali burn.

Methods: For the treatment group, 50 mice each received subconjunctival injection (5 microl) of 10(9) plaque forming units of replication-defective Ad-Vasohibin-1. Control group mice received the same dosage of blank adenoviral vector (AdNull). Five days after injection, corneal neovascularization (CNV) was induced by placing 2.5 microl of 0.1 M NaOH on the right cornea for 30 s. Subsequently, CNV was observed and photographed every 3 days for a total duration of 9 days after the alkali burn. The percentage of neovascularized area was measured and compared with the AdNull control. The expression of human vasohibin-1 protein was detected by immunohistochemistry and western blotting at 5, 8, and 14 days after injection. The mRNA expression levels of murine vascular endothelial growth factor (Vegf), VEGF receptor 1 and 2 (Vegfr1, Vegfr2), and vasohibin-1 (Vash1) were analyzed and compared by real time quantitative reverse-transcription polymerase chain reaction.

Results: The percentage of neovascularized area within the cornea was significantly reduced in mice treated with Ad-Vasohibin-1 compared to mice treated with AdNull at every time point after alkali-induced injury (7.11%+/-3.91% and 15.48%+/-1.79% of corneal area in the treatment and control groups, respectively, on day 3; 31.64%+/-4.71% and 43.93%+/-6.15% on day 6, and 45.02%+/-9.98% and 66.24%+/-7.17% on day 9, all p<0.001). Human vasohibin-1 protein was detected at the injection sites on day 3 after corneal burn and was highly expressed in the central subepithelial stroma and co-localized with neovascularized vessels within the alkali-treated cornea on day 6. On day 9, the peripheral cornea exhibited a similar staining pattern as the central cornea, but a more intense vasohibin-1 immunostaining signal was detected in the deep stroma. Some of the vasohibin-1 stain signal diffused into the frontal and deep stroma of the central cornea and was not co-localized with new vessels. By contrast, in mice injected with AdNull or normal corneas, no vasohibin-1 stain signal was detected within the corneas. Vasohibin-1 protein expression within treated corneas was also further confirmed by western blotting on day 5. Expression appeared to peak by day 8 and was maintained at high levels until day 14. However, Vasohibin-1 protein was not detected in the corneas of normal mice or mice treated with AdNull. Real-time quantitative reverse-transcription polymerase chain reaction analysis showed that expression of Vegfr2 and endogenous Vash1 mRNA were significantly decreased in the treatment versus control group (t(1)=-2.161, p(1)=0.047; t(2)=-2.236, p(2)=0.041). In contrast, there were no significant differences in Vegf and Vegfr1 mRNA expression levels between the treatment and control groups (p>0.05 for both).

Conclusions: Subconjunctival injection of Ad-Vasohibin-1 significantly reduces corneal neovascularization in alkali-treated mouse corneas. This effect of anti-neovascularization may be related to the downregulation of Vegfr2 expression.

Figure 1

The percentages of neovascularized cornea at each time point. The percentage of neovascularized cornea increased over time in both groups but was significantly reduced in mice with subconjunctival injection of recombinant adenovirus encoding human Vasohibin-1 gene (Ad-Vasohibin-1) compared with the blank adenoviral vector (AdNull) at all time points after alkali-induced injury. On day 3, there was statistical difference in the percentage of CNV area between Ad-Vasohibin-1 (n=38) and Ad Null (n=37) groups (F=50.26, t=–11.940, 95% CI of the difference: –9.78% to –6.80%). The same results were obtained on day 6 between Ad-Vasohibin-1 (n=25) and Ad Null (n=24) groups (F=3.953, t=–7.868, 95% CI of the difference: –15.42% to –9.14%) and on day 9 between Ad-Vasohibin-1 (n=22) and Ad Null (n=21) groups (F=2.318, t=–7.975, 95% CI of the difference: –26.60% to –15.85%). The asterisk indicates a p<0.01.

Figure 2

Corneal neovascularization in alkali-treated mice. The area of neovascularized cornea increased over time in both the experimental and control groups but was significantly reduced in mice treated with subconjunctival injection of Ad-Vasohibin-1 (A–C) compared to the control group with AdNull (D–F) on day 3 (A, D), day 6 (B, E), and day 9 (C, F) after alkali-induced injury.

Figure 3

The expression of exogenous human vasohibin-1 protein in mouse corneas at different time points after alkali-induced injury. Ad-Vasohibin-1 was injected subconjunctivally 5 days before alkali-induced corneal injury at a titer of 10⁹ viral particles. A–C: On day 3 after alkali-induced injury (8 days after injection), vasohibin-1 was expressed within the injection sites and some staining signals were detected at the roots of the iris. D–F:On day 6 after alkali treatment, vasohibin-1 was highly expressed in the subepithelial stroma of the cornea, and some immunostaining was detected in the deep stroma of the central cornea. Vasohibin-1 expression was highly co-localized with corneal neovascularzation. Immunostaining was diffusely distributed in the frontal stroma, which was not co-localized with new blood vessels within the cornea. G–I: On day 9, peripheral cornea showed a similar staining pattern as central cornea, but more vasohibin-1 expression was detected in the deep corneal stroma. J–L: There was no positive immunostaining for vasohibin-1 and CD31 antigen in normal cornea without subconjunctival injection.

Figure 4

Expression of vasohibin-1 protein within mouse cornea, as detected by western blot analysis. Mice were administered with subconjunctival injection of Ad-Vasohibin-1 at 10⁹ viral particles. Corneas were harvested at different time points, and protein samples were extracted. Ten corneal samples were pooled together for each western blot analysis. A mouse antihuman vasohibin-1 monoclonal antibody was used for the western blot. No vasohibin-1 western blot signal was detected before Ad-Vasohibin-1 was injected. A relatively weak signal was detected on day 5 and was maintained at high levels until day 14 after injection.

Figure 5

Quantitative comparison of gene expression between the treatment and control groups by real time RT–PCR. The transcript quantification of target genes was normalized relative to the standard housekeeping gene encoding Gapdh (arbitrary units, AU). The expression of Vegfr2 and endogenous Vash1 mRNA transcripts were significantly decreased in the treatment group compared to the control group (A, B). Meanwhile, there were slightly decreased but no significant differences in the expression levels of Vegf and Vegfr1 (C, D) in the treatment versus control group (p>0.05 for both genes).