Delayed Corneal Epithelial Wound Healing in Aquaporin–3–Null Mice

Abstract

Abstract: : Purpose: The water/glycerol–transporting protein aquaporin (AQP)–3 is expressed in corneal and conjunctival epithelia in human and mouse eye. Here, we test the hypothesis that AQP3 is involved in corneal epithelial wound healing. The motivation for this study was the previously demonstrated role of AQP3 in epidermal formation after injury, and a recently discovered role of aquaporins in cell migration. Methods: Central corneal epithelial defects were created in wild–type and AQP3–null mice in a CD1 genetic background. Wounds were demarcated with a trephine and full–thickness epithelia were removed with a blunt blade without basement membrane disruption. Centripetal advance of epithelia during healing was monitored at specified times by fluorescein imaging. Paraffin sections were also obtained at baseline and at different times after wound creation for morphology and immunohistochemistry. Wild–type AQP3 transcript expression in various ocular surface regions was analyzed by quantitative RT–PCR. In vivo ocular surface open–circuit potential differences (PDs) were recorded to assess knock–out epithelial resistance and ion transport properties. Glycerol content of homogenized full–thickness corneal epithelium was measured by a glycerol oxidase–based colorimetric assay. Results: Baseline corneal morphology and ocular PDs were similar in wildtype and AQP3–null mice, suggesting similar corneal epithelial integrity in the two groups. Corneal epithelial glycerol content was not affected by AQP3 deletion (in nmol/µg protein: 2.0 ± 0.2 versus 1.7 ± 0.1 in wildtype mice). Immunohistochemistry and RT–PCR showed strong AQP3 expression in conjunctival epithelium and at the corneal limbus, with relatively weaker expression in the central corneal epithelium. At 24 hours after wound creation, AQP3–null defects were 85 ± 4% of their original area (versus 33 ± 1% for wild–type; SE, 8 mice per group). At 48 hours, wounds in most AQP3–deficient mice remained incompletely closed (17 ± 4% area) compared to wild–type (2 ± 2%). Histology confirmed that corneas from wildtype mice largely resurfaced with stratified epithelia by 48 hours, whereas many AQP3–deficient corneas lacked even a basal epithelial cell layer and had significant inflammatory reaction in the anterior stroma. Conclusions: Our results provide evidence for a novel role for AQP3 in corneal wound healing. This defect is probably not related to baseline differences in glycerol content or to the glycerol–transporting role of AQP3. We propose that reduced water permeability in AQP3–null limbal epithelial cells reduces cell migration during healing.

Keywords: wound healing • cornea: epithelium • transgenics/knock-outs