Two Waves of Neutrophil Emigration in Response to Corneal Epithelial Abrasion: Distinct Adhesion Molecule Requirements

Abstract

purpose. Corneal abrasion results in an inflammatory response characterized by leukocyte emigration into the corneal stroma. Adhesion molecules play a critical role in leukocyte emigration to wound sites, but differences are evident in different vascular beds. In this study, the contributions of two families of adhesion molecules to neutrophil emigration into the cornea were investigated.

methods. Re-epithelialization, patterns of neutrophil influx and CXC chemokine production were assessed in C57Bl/6 mice after removal of a 2-mm diameter area of central corneal epithelium. Comparisons were made between wild-type (WT) mice and mice with targeted deletions of genes for CD18 (CD18^−/−) or P- and E-selectin (P/E-sel^−/−) or in mice with antibody-induced neutropenia.

results. Wild-type mice exhibited neutrophil emigration in two waves, the first peaking at 18 hours and the second at 30 hours after wounding, 6 hours after epithelial wound closure and peak levels of corneal CXCL1. In CD18^−/− animals, only a single wave of neutrophil influx was seen, and it was temporally and quantitatively equivalent to the second wave in WT. In P/E-sel^−/− mice, neutrophil influx was markedly depressed throughout the 48-hour observation period. Re-epithelialization was significantly delayed in mice with adhesion molecule deletions and in neutropenic animals. Transfer of wild-type leukocytes into CD18^−/− mice resulted in neutrophil emigration into the injured cornea within 18 hours of wounding and improved closure of the epithelium.

conclusions. Neutrophil emigration into corneal stroma after epithelial abrasion occurs in two waves. The first is dependent on CD18 integrins and selectins, whereas the second is CD18-independent but requires selectins. Early leukocyte emigration appears to promote re-epithelialization.

Corneal epithelium plays important roles in the maintenance of corneal function and integrity. Prolonged corneal epithelial defects may cause corneal opacity, neovascularization, bacterial infection, and visual loss. To prevent such complications, corneal epithelial wounds must be re-epithelialized as quickly as possible. The increase in refractive surgery over the past decade has expanded the importance of understanding basic mechanisms, since the wound healing response is a major determinant of efficacy and safety for refractive surgical procedures. 

As with other kinds of epithelium (e.g., mucosal and dermal), corneal epithelial healing is a complex process influenced by overlapping events such as cell proliferation and migration and the inflammatory response to injury. Superficial wounds result in leukocyte infiltration arising from limbal vessels ^{1 2} and progressing to the area of the wound. Keratocytes beneath the abraded area rapidly undergo apoptosis ³ and, within a few hours, neutrophils arrive at the region of injury by migrating through the extracellular matrix of the corneal stroma. ⁴ Under normal circumstances, re-epithelialization progresses rapidly, ⁵ the number of infiltrating neutrophils return to baseline, presumably as a result of apoptosis, ⁶ and keratocytes repopulate the stroma beneath the repair. ^{7 8} In the absence of stromal injury, there is little if any fibrotic response. ⁹  

In this article, we have begun basic studies to assess the contributions of adhesion molecules to the events just described. We analyzed the kinetics of re-epithelialization and neutrophil influx into the corneas of mice after a simple epithelial injury and determined the relative contributions of adhesion molecules in the CD18 and selectin families. These evaluations were performed in mice with a targeted deletion of CD18 or in mice with both P- and E-selectin deleted. The choice of these specific adhesion molecules for this investigation is based on the fact that these have been found to be necessary for leukocyte emigration at sites of inflammation in other regions of the body. CD18 is the β subunit of the β2 integrin family of leukocyte adhesion molecules which consists of four members, each heterodimer is composed of αβ subunits. ¹⁰ Expression of CD18 is limited to leukocytes, and the α subunit determines ligand specificity. A broad range of functions have been found for this family, and a dominant feature of the knockout phenotype is the profound reduction in neutrophil influx under most but not all inflammatory conditions. ¹¹ These integrins allow leukocytes to develop stable adhesive interactions with vascular endothelial cells before extravasation. E- and P-selectin are two members of the selectin family of adhesion molecules that are expressed on the surface of endothelial cells at sites of inflammation. These are necessary for the initial tethering of leukocytes under flow conditions, and blocking their function profoundly reduces leukocyte accumulation at sites of inflammation. ¹² Our results show that neutrophil emigration into the stroma of the cornea occurs in two waves with distinct requirements for adhesion molecules, and in the absence of these adhesion molecules, re-epithelialization is significantly delayed. 

Methods

Animals

C57Bl/6 mice were purchased from Harlan (Indianapolis, IN). CD18^−/− mice ¹³ were backcrossed at least 10 generations with C57Bl/6 mice. Mice deficient in both P- and E-selectin were developed as previously described ¹⁴ and backcrossed 12 generations with C57B1/6 mice. All mice used in this study were 6 to 8 weeks old, weighed 18 to 20 g, and were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and institutional and federal guidelines. 

Wound Model

Mice were anesthetized with intraperitoneal injection of pentobarbital sodium solution. The central corneal epithelium was demarcated with a 2-mm trephine and then removed using a diamond blade for refractive surgery (Accutome, Malvern, PA) under a dissecting microscope, as described elsewhere. ¹⁵ Care was taken not to injure the epithelial basement membrane and stroma. Assessment of wound closure used fluorescein staining every 6 hours and digital analysis of the stained area. ¹⁵ Rat anti-mouse Gr-1 monoclonal antibody (PharMingen, San Diego, CA) was used to deplete peripheral neutrophils. The antibody (0.25 mg) was administered intraperitoneally into mice 1 day before corneal epithelial wounding. Treatment with this dose of the antibody induced severe neutropenia for up to 5 days, as assessed by counting more than 200 leukocytes on blood smears, similar to that reported. ¹⁶ Control mice received an equivalent amount of normal rat IgG (PharMingen, San Diego, CA). Studies to introduce wild-type cells into CD18^−/− mice were also performed. Whole blood was acquired by aspirating the abdominal main vein of fully anesthetized and ventilated donor (C57Bl/6 wild-type) mice. Blood was immediately anticoagulated by collecting into a one-tenth volume of acid citrate dextrose (ACD) solution (Sigma-Aldrich, St. Louis, MO), with immediate mixing on collection of the necessary volume. Leukocyte-rich plasma was acquired by centrifuging the whole blood at low speed (260g, 8 minutes, 22°C) and was used for intravenous transfusion within 1 hour after preparation. A 0.2-mL volume of leukocytes infused slowly through a tail vein after corneal epithelial wounding. 

Immunohistology

Wounded corneas were dissected, taking care to include the limbus; fixed; permeabilized; and incubated with labeled monoclonal antibodies as described elsewhere. ¹⁵ Anti-Gr-1-FITC and anti-CD31/PECAM-1-PE (PharMingen) antibodies were selected for neutrophils and limbal vessels separately. Radial cuts were made in the cornea so that it could be flattened by a coverslip, and it was mounted (Airvol; Air Products and Chemicals, Allentown, PA) containing 1 μM 4′,6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich), to assess nuclear morphology and cell division. Digital images were captured and saved for digital analysis (Delta Vision; Applied Precision, Issaquah, WA). To compare the relative level of neutrophils in the different areas from the limbus to the central cornea, we counted neutrophils in each cornea separately. At least four corneas were examined for immunohistology and four quadrants were analyzed for each, to obtain the average number per field. The limbus was defined as the intervening zone between the cornea and conjunctiva as the most peripheral field. 

Determination of Chemokines

CXCL1 (KC) and CXCL5 (LIX) were analyzed by ELISA (R&D Systems, Inc., Minneapolis, MN) in extracts of corneas collected at 6-hour intervals throughout the observation period. Four corneas at each time point were homogenized in 600 μL of phosphate-buffered saline and subjected to three freeze-thaw cycles and sonication for 30 s. All homogenates were frozen at −70°C until they were used in the assay. 

Statistical Analysis

Data analysis was performed using two-way repeated-measures ANOVA and pair-wise multiple comparisons using the Tukey test. P <0.05 was considered significant. Data are expressed as the mean ± SEM. 

Results

The basic model of corneal wound injury involved removal of a 2-mm diameter area of epithelium from the center of the cornea, taking care to damage the underlying stroma as little as possible. The process of re-epithelialization monitored by fluorescein staining of the exposed stroma was found to be essentially complete by 24 hours after wounding in wild-type C57Bl/6 mice (Fig. 1) . Mice that were deficient in CD18, neutropenic, or deficient in both E- and P-selectin exhibited delays of 6 to 18 hours in wound healing. 

Wild-Type Mice

The kinetics of wound closure in wild-type mice is shown in Figure 2 . Epithelial cell division was detected within the uninjured epithelium in the periphery of the cornea at 12 hours and peaked between 18 and 36 hours after wounding. The epithelial filling in the wound exhibited some division after the wound was covered at 30 hours. Restoration of the full density of epithelium was complete by 96 hours (see Fig. 4B ). 

Neutrophils (Gr-1+ cells with DAPI-stained multilobed nuclei) were observed in the stroma of the cornea within 2 hours. Their extravascular position was revealed by staining vessels with anti-CD31 antibody. As shown in Figure 3A , the accumulation of neutrophils in the corneal stroma exhibited two peaks, one at 12 to 18 hours, and another at 30 to 36 hours. By 48 hours, neutrophil levels within the stroma were near those found before injury. The levels of neutrophils at 24 hours after wounding were significantly lower than either peak (Fig. 3A) . Analysis of the distribution of neutrophils across the diameter of the cornea using the pattern for microscopic analysis shown in Figure 2 , revealed significant differences between peaks 1 and 2 (Fig. 3B) . Significantly more neutrophils were present in the stroma beneath the wound at the 18-hour time point than at the 24- and 30-hour time points (Fig. 3B) . It appears that the second peak of neutrophil accumulation was predominately in the periphery near the limbus (zone 1). In addition, comparing the distributions of neutrophils at 18 and 24 hours reveals that neutrophil levels in all zones of the cornea decreased at 24 hours (Fig. 3B) . 

The relationship of wound closure to the two peaks of neutrophil accumulation is shown in Figure 3C . It is evident that peak 1 occurs before wound closure, and peak 2 occurs after wound closure. In addition, Figure 3Cdemonstrates that the two peaks are evident in the limbus, the vascularized area adjacent to the cornea. Since we were able to see neutrophils within the immediate vicinity of the limbal vessels within 2 hours of wounding, it appears that the primary origin of neutrophils within the stroma was emigration from the limbal vessels. The CXC chemokines CXCL1 (KC) and CXCL5 (LIX) were recoverable from extracts of injured corneas and demonstrated elevations during the period of wound closure, returning to baseline after 24 to 30 hours (Fig. 3D) . CXCL2 (MIP-2) was also detected, but at very low levels (data not shown). 

CD18^−/− Mice

As indicated in Figure 1 , delayed wound closure was significant in the CD18^−/− mice (Figs. 1 4A) . Analysis of neutrophil accumulation over the period of the two waves of accumulation in the wild-type mice revealed only one wave of neutrophils in mice missing the CD18 integrins (Fig. 4A) . This wave corresponded temporally to the second wave in the wild-type mice and is evident visually in Figure 5 . The absence of neutrophils in the limbus during the 12- to 18-hour period was consistent throughout the cornea, and there were no neutrophils detected in the central corneal stroma beneath the abraded epithelium (Fig. 4A) . However, at 24 hours, when neutrophils began emigrating, they were found throughout the entire cornea from the limbus to the center of the wound, indicating that not only was CD18 unnecessary for emigration at this time, but it was also unnecessary for migration through the avascular stroma of the cornea. As evidenced in Figure 4C , elevations in CXC chemokines also preceded the 30- to 36-hour peak of neutrophils in the CD18^−/− corneas (as was seen in the wild-type mice), but in contrast to the wild-type, the levels of the chemokines were approximately nine times higher. 

Also shown in Figure 4Ais the kinetics of wound closure, indicating a 6-hour delay in re-epithelialization of the wound in the CD18^−/− mice. The delay in healing was further revealed by analysis of basal epithelial cell density. At 96 hours after wounding, when wild-type epithelial density was back to normal levels, that in CD18^−/− mice was significantly reduced (Fig. 4B) . 

In an effort to determine whether early leukocyte influx could be reinstated in CD18^−/− knockout mice given intravenous injections of wild-type leukocytes, corneas were examined at 18 hours after injury. A significant increase in neutrophils within the corneas (n = 4) was found (Fig. 4D) . In contrast to 18-hour wounded corneas in CD18^−/− mice, transferred wild-type neutrophils in CD18^−/− mice were found at the wound site (mean ± SEM for CD18^−/− mice, 1.2 ± 1.1 cells per field in zones 3, 4, and 5 of Figure 2 ; for CD18^−/− with transferred wild-type leukocytes, 24.6 ± 14.2; for WT, 49.2 ± 13.1). Most of these leukocytes within the wound region were wild-type as indicated by CD11b positivity (Fig. 6) . Using epithelial basal cell density as a sensitive measure of wound closure, CD18^−/− mice with transferred wild-type leukocytes had significantly more epithelial cells in the center of the wound than CD18^−/− mice (Fig. 4B) . 

E- and P-selectin Double-Knockout Mice

The contributions of selectins are reflected in the response of mice with targeted deletions of both E- and P-selectins. The neutrophil accumulation in the corneas after wounding revealed a pattern very different from that in either the wild-type or the CD18^−/− mice. Although there was some neutrophil accumulation during the 12- to 18-hour period after wounding, it was significantly lower than that in the wild-type (Fig. 7A) . The level of neutrophils throughout the cornea at all time points studied was markedly reduced. CXCL1 and CXCL5 levels extracted from whole corneas was correspondingly low in these mice (Fig. 7B)when compared with those of wild-type mice (Fig. 3D) . Wound healing in these mice was significantly delayed until 18 hours after wounding, and wound closure was not evident until 42 hours (Fig. 7C) . Epithelial density was significantly delayed at 96 hours after wounding (Fig. 7D) . 

Neutropenic Mice

Given the observations with the adhesion molecule-knockout mice that neutrophil influx was markedly reduced and that epithelial closure was delayed, we performed an experiment to assess the effects of preventing leukocyte influx in wild-type mice by inducing profound neutropenia over the duration of the primary healing period (i.e., approximately 30 hours). Re-epithelialization as revealed by fluorescence staining of corneal wounds was delayed by more than 24 hours (Figs. 1 8) . 

Discussion

In an effort to begin analyzing the mechanisms relevant to re-epithelialization in the wounded cornea that are disrupted by the absence of adhesion molecules, we began a detailed analysis of the migration patterns of neutrophils. In the current study, the results provided evidence that simple corneal epithelial injury resulted in two waves of neutrophil migration into the corneal stroma. The first wave peaked at 12 to 18 hours and the second wave between 30 and 36 hours. Both waves of accumulation of neutrophils were evident in the region of the limbal vessels beneath the uninjured epithelium, whereas accumulation in the stroma beneath the epithelial injury seemed predominately associated with the first wave. Analysis of mice deficient in CD18 showed inhibition of the first wave. This was not surprising, given observations of acute inflammation in other tissues in mice and other species deficient in CD18. ¹¹ CD18 is the β subunit of this family (β2) of integrins. Its absence results in failure of expression of all four members of the family and is associated with a very substantial peripheral blood leukocytosis. ¹¹ The failure of any detected emigration within the period of the wild-type first wave clearly indicates a requirement of this integrin for neutrophil emigration through the limbal vessels. The finding of a single wave of neutrophils in CD18^−/− mice corresponding temporally to the second wild-type wave was unexpected, however. It occurred after closure of the epithelial wound and approximately 6 hours after the peak of CXCL1 levels. The extension of this wave into the center of the cornea and the 10-fold elevation in chemokines indicate mechanistic differences. This is the first observation of a CD18-independent emigration of neutrophils into the cornea and the first observation of a vascular bed where both CD18-dependent and -independent pathways occur in response to a single stimulus. 

Earlier studies of wound healing in the cornea have provided data that may be relevant to an interpretation of the data in the current manuscript. Zhu and Dana ⁴ investigating a more severe injury (11-0 sutures within the corneal stroma) than simple corneal abrasion found that limbal vessels in BALB/c mice had increased expression of intracellular adhesion molecule (ICAM)-1 by 8 hours after wounding and that expression remained high throughout a 14-day observation period. In addition, they found that C57Bl/6 mice (the strain used in the current studies) with a targeted deletion of domain 4 of ICAM-1 exhibited ∼50% reductions in neutrophil influx at all observation times after wounding. Since ICAM-1 is one of the ligands for the family of CD18 integrins, ¹⁷ its expression in the limbal vessels within hours of wounding is consistent with a role for CD18 integrins in the early wave of neutrophil extravasation. Schultz et al. ¹⁸ found that systemic administration of anti-CD18 monoclonal antibody IB4 to rabbits markedly reduced early leukocyte influx into abraded corneas. The mechanism accounting for the CD18-independent wave of neutrophils remains to be defined. 

Initially, the CD18-independent adhesion pathway for neutrophils was thought to be present only in the lungs, perhaps because of the unique structure of the pulmonary capillary bed and the unique cell types present in the lungs. ¹⁹ However, there is also a CD18-independent neutrophil migration into organs such as kidney, liver, heart, and intestinal mucosa. ^{20 21 22 23} The molecular mechanisms for this pathway are not understood, though multiple lines of evidence suggest that this pathway is influenced by the type of inflammatory stimulus, the specific adhesion molecules expressed, and the chemoattractant expressed. Neutrophil response to Escherichia coli endotoxin, certain Gram-negative organisms, immune complexes, and IL-1β occur through pathways that require CD18 integrins. In contrast, hyperoxia, hydrochloric acid, Streptococcus pneumoniae, Group B Streptococcus, and Staphylococcus aureus induce emigration by CD18-independent pathways. ^{24 25 26 27 28 29 30} Bowden et al. ²² found that vascular cell adhesion molecule (VCAM)-1 expression in the vessels of the heart after ischemia and reperfusion supported CD18-independent migration. Zhu and Dana ⁴ were unable to find VCAM-1 expression in limbal vessels in their murine corneal wound model possibly indicating that this pathway may not account for the CD18-independent migration in the present studies. 

Chemotactic factors seem to stimulate CD18-independent transendothelial migration of neutrophils, ^{31 32 33} and this may be of significance in our current model, given the observation that CXC chemokine levels in the corneas peaked at 6 hours before the peak of the CD18-independent wave of neutrophil emigration. CXCL1 and CXCL5 are produced by interstitial cells in the corneal stroma after injury ^{34 35} along with cytokines ³⁶ that may contribute to the inflammatory events (e.g., G-CSF). Our observations that the peak of CXCL5 (LIX) corresponded to the peak of neutrophil influx in both the wild-type and CD18^−/− mice is of interest, given the recent finding that G-CSF can induce the expression of CXCL5 by neutrophils. ³⁷ Explanations for the marked differences in CXCL1 levels in the CD18^−/− and selectin-deficient mice are not evident in this study. Further studies are needed, to define the specific sources of the chemokines to determine which cell types are responsible, and to determine whether emigrating neutrophils serve as sources of the chemokine. 

In this report, we provide evidence that CD18 integrins and endothelial selectins are necessary for efficient re-epithelialization of the cornea in mice after mechanical abrasion. These results are consistent with those of Gan et al. ³⁸ They found delayed wound closure in corneas of rabbits treated with fucoidin, an agent that has been reported to block the function of selectins and thereby the emigration of leukocytes. In the current studies, mice deficient in the endothelial selectins exhibited significant delays in corneal epithelial wound healing and marked reductions in neutrophil emigration in all regions of the cornea throughout the entire observational period. These results are consistent with observations of dermal wound healing. In unpublished studies, we have found that 4-mm punch biopsy wounds in CD18^−/− mice exhibit significant delays in re-epithelialization, and observations of wounds in patients with CD18 deficiency (i.e., the LAD 1 syndrome ¹⁰ ) show poor healing. Subramaniam et al. ³⁹ analyzed dermal wounds in the P/E-selectin-deficient mice and found significant delays in re-epithelialization. Given that the response of most blood leukocytes would be influenced by the absence of these adhesion molecules, it is impossible to know what aspect of the wound-healing mechanisms are adversely affected in the knockout animals. 

Regarding the contributions of leukocytes to corneal re-epithelialization, two experiments in this study support a positive contribution to healing. The first is the observation that transfer of wild-type leukocytes intravenously into CD18^−/− mice partially restored neutrophil emigration into the corneal wound area and significantly increased the epithelial coverage of the abraded area at 18 hours. The partial effect of this transfer experiment may be linked to the inability to restore the full level of neutrophil emigration. Technically, it has not been possible to isolate neutrophils without some degree of activation, thus inducing abnormal distribution of transferred cells (e.g., excessive sequestration within the liver and bone marrow ⁴⁰ ). We chose to collect whole leukocyte preparations without neutrophil isolation, because this resulted in lower levels of activation (assessed for shedding of L-selectin, an adhesion molecule rapidly cleaved from the cell surface with activation; data not shown), but there is still concern that this only partially restores normal efficiency of acute inflammation. Thus, an additional approach involving neutropenic animals to assess wound closure was used. The results of the experiments are consistent with the interpretation that lack of neutrophil influx (i.e., with neutropenia, CD18 deficiency, or selectin deficiency) delays wound closure, whereas partial restoration of neutrophil influx with wild-type neutrophils transferred into CD18-deficient mice enhances wound closure. In addition, observations in the CD18-deficient mice revealed a shorter period of delay in wound closure than in mice with either prolonged neutropenia or in selectin-deficient animals with profound inhibition of neutrophil influx. The CD18-independent neutrophil emigration in CD18^−/− mice began at approximately 24 hours after wounding, and epithelial closure occurred shortly thereafter. In contrast, wound closure in the selectin-deficient or profoundly neutropenic mice was further delayed. It is important, however, to emphasize that while reductions in neutrophil emigration reflect the contributions of adhesion molecules and correlate with delays in healing, the mobility of many other types of leukocytes is also dependent on these adhesion molecules. In addition, the mobility of other leukocytes may also be influenced by the emigration of neutrophils. Thus, it is premature to conclude that neutrophils are directly responsible for the apparent benefit of early inflammation after corneal abrasion. 

 Supported by National Aeronautics and Space Administration (NASA) Grant NAG 9-1253; National Heart, Lung, and Blood Institute Grant HL070537; National Institute of Allergy and Infectious Diseases Grant AI46773; and National Natural Science Foundation of China Grant 39970250.

 Submitted for publication September 7, 2005; revised December 16, 2005; accepted March 6, 2006.

 Disclosure: Z. Li, None; A.R. Burns, None; C.W. Smith, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Fluorescein staining of abraded corneas revealing wound closure delays in the mice with targeted deletions of CD18 or P- and E-selectin. The results are representative of at least five mice of each strain. The hours indicate the time after the abrasion.

Figure 2.

 

Wound closure in wild-type C57Bl/6 mice after a 2-mm diameter epithelial debridement. Left: wound closure kinetics indicating that closure was essentially complete at 24 hours (mean ± SEM, n = 10). Middle: morphometric analysis showing patterns of dividing cells across the cornea in nine ×40 fields. Right: Plotting the average number of dividing cells in the zones indicated demonstrates that epithelial cell division was highest in the outer zones 1 to 3 between 18 and 36 hours after wounding, and cell division in the center of the wound area after re-epithelialization began at 30 to 36 hours after wounding.

Figure 4.

 

Corneal response to epithelial abrasion in CD18^−/− mice. After the removal of a 2-mm diameter area of epithelium, corneas were collected for analysis at 6-hour intervals up to 48 hours. (A) Top: shows the wound-closure kinetics comparing the matched wild-type with the CD18^−/− mice. *P < 0.01. Wholemount preparations were stained with anti-Gr-1-FITC and DAPI, to allow analysis of neutrophil emigration. Neutrophils were counted in zones 1 through 5 (see Fig. 2 ) in four quadrants of each cornea, and the sum of these counts was plotted against the time after wounding (bottom). (B) Basal epithelial cell density determined across the mid diameter of the cornea 24 and 96 hours after wounding. The data compare wild-type (filled diamonds) and CD18^−/− (open squares) mice showing significant delays in the recovery of epithelial density in the CD18^−/− mice in all regions of the cornea (**P < 0.01). In addition, the top graph includes data from CD18^−/− mice into which wild-type leukocytes have been transferred (stippled triangle; **P < 0.01). (C) CXC chemokines extracted from CD18^−/− corneas after epithelial abrasion. The plot of neutrophils is the same data set as in (A) and in contrast to the plots in (A), the total neutrophil counts in zones 1 to 5 are given. Note that CXCL1 (KC) exhibited a major peak of 1114 pg per cornea at 6 hours before the peak of neutrophil influx at 30 hours, whereas CXCL5 (LIX) had a single wave peaking at 300 pg per cornea and coincident with the wave of neutrophils at 30 hours. (D) Leukocyte influx into corneas of CD18^−/− mice after intravenous transfer of wild-type leukocytes. CD18^−/− mice received three tail vein injections of wild-type leukocytes at 4-hour intervals beginning immediately after corneal abrasion. Corneas were collected at 18 hours after wounding and wholemounts were processed for analysis of Gr-1- and CD11b-positive cells by microscopic analysis of immunofluorescence. Neutrophils were counted in zones 1 to 5 (see Fig. 2 ) in four quadrants of each cornea, and the sum of these counts was plotted. Results compare CD18^−/− mice receiving neutrophils with the control (wild-type and CD18^−/−) mice.

Figure 3.

 

Corneal response to epithelial abrasion in wild-type mice. After the removal of a 2-mm diameter area of epithelium, corneas were collected for analysis at 6-hour intervals up to 48 hours. (A) Wholemount preparations were stained with anti-Gr-1-FITC and DAPI, to allow analysis of neutrophil emigration. Neutrophils were counted in zones 1 to 5 (see Fig. 2 ) in four quadrants of each cornea, and the sum of these counts was plotted against the time after wounding. *P < 0.01, n = 8. (B) The distribution of neutrophils across the mid-diameter of corneas is plotted for 18, 24, and 30 hours after wounding. Note that the level of neutrophils in the central region of the cornea (i.e., zones 3, 4, and 5) is significantly (*P < 0.01) greater at 18 hours than at 24 or 30 hours. Also the level of neutrophils in the limbus (zone 1) is significantly greater at 18 and 30 hours after wounding than at 24 hours. (C) Kinetic patterns for neutrophil influx and wound closure. Note that the biphasic pattern of neutrophil influx at the limbus was not seen in the center of the cornea. Also, note that the second wave of neutrophils in limbus occurred after epithelial closure. (D) CXC chemokines extracted from corneas after epithelial abrasion. The plot of neutrophils is a subset of the data as in (A) and is included for reference. Note that CXCL1 (KC) exhibited a biphasic pattern preceding that of the neutrophils, peaking at 174 pg per cornea at 6 hours before the second neutrophil wave, whereas CXCL5 (LIX) had a single wave peaking at 50 pg per cornea, corresponding to that of the first wave of neutrophils.

Figure 5.

 

Fluorescence microscopy images of the limbal region of corneas after epithelial wounding. After the removal of a 2-mm diameter area of epithelium, corneas were collected for analysis at 6-hour intervals. Wholemount preparations were stained with anti-Gr-1-FITC and DAPI, to allow analysis of neutrophil emigration, and with anti-CD31-PE to allow recognition of the microvasculature of the limbus.

Figure 6.

 

Leukocyte influx into corneas of CD18^−/− mice after intravenous transfer of wild-type leukocytes. CD18^−/− Mice received three tail vein injections of wild-type leukocytes at 4-hour intervals beginning immediately after corneal abrasion. Corneas were collected at 18 hours after wounding and wholemounts were processed for analysis of Gr-1- and CD11b-positive cells by microscopic analysis of immunofluorescence. Micrographs of the central wound area showing that in wild-type control (WT) mice, neutrophils were abundant and positive for CD11b; in CD18^−/− control animals, neutrophils were rare and none of the cells were positive for CD11b. In CD18^−/− mice receiving transfusions of wild-type leukocytes (TL), neutrophils were evident and positive for CD11b.

Figure 7.

 

Corneal response to epithelial abrasion in mice deficient in both P- and E-selectin mice. After the removal of a 2-mm diameter area of epithelium, corneas were collected for analysis at 6-hour intervals up to 48 hours. Wholemount preparations were stained with anti-Gr-1-FITC and DAPI, to allow analysis of neutrophil emigration. (A) Neutrophils were counted in zones 1 and 2 (see Fig. 2 ) in four quadrants of each cornea, and the sum of these counts was plotted against the time after wounding. (B) CXC chemokines extracted from corneas after epithelial abrasion. (C) Wound closure kinetics showing significant delays in the knockout mice (*P < 0.01). (D) Basal epithelial density across the mid-diameter of the cornea at 96 hours after wounding showing significant delays in the knockout mice.

Figure 8.

 

Corneal response in wild-type neutropenic mice after epithelial abrasion. Fluorescein staining of abraded corneas (n = 4) was quantified at 4-hour intervals and plotted as the percentage of open area. *P < 0.05; **P < 0.01.

The authors thank Elizabeth Priest for providing assistance.