Partial denervation of sub-basal axons persists following debridement wounds to the mouse cornea

Abstract

Although sensory reinnervation occurs after injury in the peripheral nervous system, poor reinnervation in the elderly and those with diabetes often leads to pathology. Here we quantify sub-basal axon density in the central and peripheral mouse cornea over time after three different types of injury. The mouse cornea is highly innervated with a dense array of sub-basal nerves that form a spiral called the vortex at the corneal center or apex; these nerves are readily detected within flat mounted corneas. After anesthesia, corneal epithelial cells were removed using either a dulled blade or a rotating burr within an area demarcated centrally with a 1.5 mm trephine. A third wound type, superficial trephination, involved demarcating the area with the 1.5 mm trephine but not removing cells. By 7 days after superficial trephination, sub-basal axon density returns to control levels; by 28 days the vortex reforms. Although axon density is similar to control 14 days after dulled blade and rotating burr wounding, defects in axon morphology at the corneal apex remain. After 14 days, axons retract from the center leaving the sub-basal axon density reduced by 37.2 and 36.8% at 28 days after dulled blade and rotating burr wounding, respectively, compared with control. Assessment of inflammation using flow cytometry shows that persistent inflammation is not a factor in the incomplete reinnervation. Expression of mRNAs encoding 22 regeneration-associated genes involved in axon targeting assessed by QPCR reveals that netrin-1 and ephrin signaling are altered after wounding. Subpopulations of corneal epithelial basal cells at the corneal apex stop expressing ki67 as early as 7 days after injury and by 14 and 28 days after wounding, many of these basal cells undergo apoptosis and die. Although sub-basal axons are restored to their normal density and morphology after superficial trephination, sub-basal axon recovery is partial after debridement wounds. The increase in corneal epithelial basal cell apoptosis at the apex observed at 14 days after corneal debridement may destabilize newly reinnervated sub-basal axons and lead to their retraction toward the periphery.

Figure 1. Subbasal axons fully reinnervate the cornea and reform the vortex over time after trephine wounding but not after dulled blade and rotating burr wounding

Representative stitched images of subbasal axons stained with βIII tubulin were taken by spinning disk confocal microscope to show the central 2 mm area of the mouse cornea encompassing the site demarcated by the 1.5 mm trephine in unwounded and corneas 1d, 2d, 3d, 4d, 7d, 14d, and 28d after wounding. The numbers of corneas used for quantification of reinnervation by Sholl analysis are indicated in the lower right corner of each image (Mag bar=500μm).

Figure 2. Fewer axons are present at the center of the cornea 28d after dulled blade and rotating burr wounds

The density of axons at the center and periphery of the cornea was determined by Sholl analysis. Data are expressed as the number of intersections present at the corneal center (left) and the periphery (right) at 1d, 2d, 3d, 4d, 7d, 14d, and 28d after wounding relative to the numbers of intersections in unwounded control corneas. Scatter plots with trendlines are presented for trephine (blue), dulled blade (red) and rotating burr (green) wounded corneas; filled circles indicate data from individual corneas. The parallel grey line at the value of 1.0 indicates axon density in control. Data are compared using one-way ANOVA and p values less than 0.05 are considered significant. The graph at the bottom shows data from all three wound types plotted together to better visualize comparisons between wound types.

Figure 3. Disruption in the epithelial surface of rotating burr wounded corneas is evident at 14d after wounding

Representative stitched images of rotating burr wounded corneas at 4d, 14d, and 28d after wounding were stained with βIII tubulin, Ki67, and DAPI and imaged by spinning disk confocal microscope. White arrows indicate surface irregularities by 14d and 28d after wounding and white asterisks points out the thickening of the epithelium seen at 28d. The white arrowheads indicate reinnervated areas, which lack Ki67 (Mag bar=100μm).

Figure 4. Expression of mRNA for Ntn1 is decreased and Efna5 is increased in corneal buttons 28d after dulled blade wounding

QPCR analysis was performed on corneal buttons at 4d, 7d, 14d, and 28d after dulled blade wounding, using 22 primers as shown here and in the Supplemental Figure 2, and the expression level for each mRNA was normalized relative to GAPDH. The asterisks above the bars indicates significant differences between time points; asterisks within bars indicate significant differences relative to control. A. Unc5b is transiently but significantly decreased at 4d. While Dcc increased significantly between 4 and 14 days, the changes were not significant relative to control. Ntn1 is reduced significantly at 28d after wounding relative to both control and to day 7. B. Efna5 is increased at 14d and 28d after wounding as compared to control and at 28 days relative to 4 days.

Figure 5. Flow cytometry shows that differences in leukocyte recruitment are seen after dulled blade and rotating burr wounds

A. Flow cytometric characterization of inflammatory cell populations in control, rotating burr, and dulled blade wounded corneas was performed at 0.75d (18 hours), 3d, 7d, 14d, and 28d after wounding. Four corneas were used per time point tested and the numbers of independent samples assayed was 4 for each wound type and time point. Representative flow cytometric data show PMNs (Ly6c+Ly6g-) and monocyte (Ly6c+Ly6g+) populations after gating for CD45+ cells. B. Leukocytes were also assessed for their expression of Ly6C and F4/80 after gating for CD45+CD11b+Ly6C- cells to determine whether there were differences in maturation of monocytes into macrophages after both types of wounding. C. The means +/− SEMs of the percentages of CD45+ cells that are PMNs, monocytes, or macrophages were determined. Asterisks above bars indicate significant differences (p values less than 0.05 by one way ANOVA) between the means of rotating burr and dulled blade. Asterisks on bars indicate significant differences compared to unwounded control.

Figure 6. Apoptotic and dead cells develop on the epithelium after corneal wounding

A. To determine when cells at the corneal center begin to undergo apoptosis, dulled blade wounded corneas were stained with γH2AX, βIII tubulin, and DAPI at 7d, 14d, and 28d after wounding. White arrow indicates patches of cells expressing reduced levels of γH2AX (γH2AXlo) at 7d after wounding. These sites remain 14d and 28d after wounding corresponding to sites lacking nerves. Arrowheads show cells expressing high levels of γH2AX suggestive that they are undergoing apoptosis at 14d and 28d after wounding (Mag bar=100μm). B. The 14 day dulled blade wounded cornea shown in A. was subjected to high resolution 3D confocal imaging. On the left is the en face view at the center of the cornea after wounding. On the right are two different 3D views tilted at different angles to show that while most of the γH2AX bright cells are in the basal cell layer, some cells are also seen in the suprabasal cell layers. (Mag bar=10μm). C. To determine whether apoptotic cells undergo apoptosis and/or die, unfixed corneas obtained immediately after sacrifice of mice 28d after dulled blade wounding were stained with Hoechst and PI and images were taken at 10× (left) and 40× (right) with a fluorescent microscope. Hoechst+ apoptotic and PI+ dead cells are present within the corneal epithelium adjacent and within sites that appear to be erosions. Asterisks show the areas in the 10× images where the 40× images were obtained.