Visualization and Characterization of Inflammatory Cell Recruitment and Migration through the Corneal Stroma in Endotoxin-Induced Keratitis

Abstract

purpose. The infiltration of inflammatory cells into the cornea is a major determinant in the outcome of keratitis. The purpose of this study was to use enhanced green fluorescence protein (EGFP) bone marrow chimeric mice to visualize and characterize the inflammatory cells that migrate to the corneal stroma during endotoxin-induced keratitis and to explore the mechanisms underlying this process.

methods. Keratitis was induced by injecting endotoxin into the corneas of EGFP chimeric mice. In vivo fluorescence stereomicroscopy was used to visualize in real time the recruitment of EGFP-positive cells at different time points. Immunohistochemistry and three-dimensional (3D) confocal analysis of whole-mount corneas was used for histologic characterization. Macrophage inflammatory protein-2 (MIP-2) chemokine was neutralized in vivo to determine its contribution to this process.

results. Recruitment of EGFP-positive inflammatory cells in the corneal stroma can be detected in vivo by 6 hours after the injection of endotoxin, and these were mainly neutrophils. Full-thickness whole corneal mount confocal image analysis showed a distinct pattern of migration of EGFP inflammatory cells through the anterior corneal stroma. Moreover, inflammatory cells did not colocalize with the injected lipopolysaccharide (LPS) deposits in the stroma but moved from all directions toward LPS, partially in response to the production of the chemokine MIP-2.

conclusions. EGFP chimeric mice and ex vivo 3D analysis of whole-mount corneas provides unique information on the interaction of infiltrating inflammatory cells in the cornea. These findings demonstrate that a chemotactic gradient triggered in part by MIP-2 is responsible for directing inflammatory cell migration through the corneal stroma.

Corneal inflammation leads to a significant loss of vision secondary to scarring and perforation caused by microbial keratitis or bacterial products, such as lipopolysaccharide (LPS). ^{1 2} Animal models of infectious keratitis and LPS or endotoxin-induced keratitis have been extremely important in determining the important role of neutrophils, chemokines, Toll-like receptors, and adhesion molecule signals involved in this process. ^{3 4 5 6 7} Although these studies have significantly helped us understand the mechanisms involved in the recruitment of inflammatory cells to the cornea, little is known about the kinetic and dynamic interactions between the inflammatory cells in the corneal environment once these inflammatory cells are recruited and leave the limbal vessels. ⁸  

The development of in vivo animal models to track and characterize the behavior of immune cells has provided key new insights into the characteristics and mechanisms of immune and inflammatory responses. ^{9 10} A major limitation of many systems is the inability to directly visualize the interaction of immune cells with the extracellular matrices or the recruitment, migration, and fate of these cells within tissue microenvironments. Recently, high-performance stereoscopic, confocal, and two-photon fluorescence microscopy has been developed to detect inflammatory responses in lymphoid and peripheral tissue. ^{11 12 13} Although in most tissues real-time observation of immune cells in live animals using these technologies requires invasive surgery, the eye has the unique advantage of providing a direct window to detect and monitor immune responses in vivo. In fact, the use of intravital microscopy to visualize immune cells in the anterior chamber has shed provocative and useful information on the dynamic process of immune regulation in the eye. ^{14 15 16}  

Similar to the anterior chamber, the translucent nature of the cornea permits in vivo visualization of inflammatory events in a noninvasive manner. Because the cornea is infiltrated by cellular mediators of inflammation during immunologic response, its use to monitor in real time the migration, differentiation, and fate of inflammatory cells has great potential. To explore the features and mechanisms of inflammatory responses in the cornea, we have developed new methods to visualize the migration and fate of inflammatory cells based on high-performance fluorescence microscopy. In vivo real-time intravital microscopy and ex vivo confocal analysis with three-dimensional (3D) reconstruction of whole-mount corneas from enhanced green fluorescence protein (EGFP) chimeric mice was performed. Bone marrow–derived inflammatory cells from these mice express EGFP, which enables the monitoring and tracking of their migration to the cornea during the development of endotoxin-induced keratitis. In this study, we characterize for the first time the route and the pattern of migration of inflammatory cells through the corneal stroma in response to an inflammatory stimulus, and we distinguish the localization of the inflammatory stimulus and the responding cells. Our results confirm the central role of neutrophils in keratitis and demonstrate that chemotactic signals are responsible for controlling their migration pattern through the corneal stroma. 

Materials and Methods

Animals

C57/BL6 and C57BL/6-TgN(ACTbEGFP)1Osb (The Jackson Laboratory, Bar Harbor, ME) 8 to 10 weeks of age were used in these experiments. ¹⁷ Animals were treated in accordance with the guidelines provided in the Association for Research in Vision and Ophthalmology Resolution on the Use of Animals in Ophthalmic and Vision Research. 

Generation of EGFP Chimeric Mice

EGFP chimeric mice were generated by injecting intravenously 2 to 5 × 10⁶ bone marrow cells from the C57BL/6-TgN (ACTbEGFP; Jackson Laboratory) into C57BL/6 recipient mice that received 12 Gy whole body radiation in 2 doses 3 hours apart, as previously described. ^{17 18 19} Two weeks after bone marrow transplantation, bone marrow, peripheral blood, and spleen cells were harvested to measure EGFP by fluorescence-activated cell sorter (FACS) analysis (FACScan; Becton Dickinson, San Jose, CA). 

Induction of Lipopolysaccharide Keratitis

Intrastromal injection of 2 mL of 2 μg LPS (Pseudomonas aeruginosa) dissolved in PBS (Sigma) or PBS only was performed in sedated animals at the limbus and paracentral cornea, as previously described. ^{20 21} Briefly, a small tunnel from the corneal epithelium to the anterior stroma was created using a 33-gauge needle (Hamilton Company, Reno, NV). Another 33-gauge needle attached to a 10-μL Hamilton syringe was passed through the tunnel into the stroma for the injection. AlexaFluor 594–conjugated LPS (Escherichia coli) (red LPS; Molecular Probes, Eugene, OR) was used to determine the distribution and location of the LPS intrastromal injection. 

In Vivo Real-Time Imaging of Immune Responses in the Corneas in EGFP Chimeric Mice

In vivo imaging of EGFP-positive inflammatory cell migration in the cornea was performed using an MZFLIII high-resolution stereo fluorescence microscope with vertical fluorescence illuminator (Leica Microsystems Inc., Bannockburn, IL). The head of each sedated EGFP-chimeric mouse was immobilized with a three-point stereotactic mouse restrainer, and images of the same animals were captured at different time points with a digital camera (SpotCam RT KE; Sterling Heights, MI). Each whole cornea was focused, and images were obtained at a magnification of 32× using a standard exposure time of 1 to 1.5 seconds. EGFP expression was quantitated (Image Pro Plus; Media Cybernetics Inc., Carlsbad, CA), and mean fluorescence density of corneas at specific time points was calculated using a consistent region of interest (ROI). 

Immunohistochemical Staining of Frozen Cornea Sections

Eyes were enucleated 24 hours after intrastromal injection, embedded, and snap frozen in optimum temperature cutting compound (Tissue-Tek; Miles Scientific, Napierville, IL). Five-micrometer frozen sections of the central cornea were generated and stained for neutrophils and macrophages with anti–NIMP-R14 (Serotec, Oxford, UK) and anti-F4/80 (Serotec), respectively, and were detected using a secondary stain with (green) goat anti-rat immunoglobulin G (IgG; AlexaFluor 488; Molecular Probes). NIMP-R14 recognizes an undetermined structure on the neutrophil plasma membrane and has been successfully used to identify neutrophils in murine corneas. ⁸ Because the endogenous EGFP in the bone marrow inflammatory cells could not be readily detected in the 5-μm sections, EGFP immunostaining was performed using a (red) directly conjugated anti-EGFP rabbit IgG fraction (AlexaFluor 594; Molecular Probes). Positively stained cells were quantified using digital software (Image Pro Plus; Media Cybernetics Inc., Silver Spring, MD). Statistical analysis (t test) was performed on three central cornea sections per group. 

Corneal Whole-Mount Confocal Imaging and Three-Dimensional Reconstruction Analysis

Corneas were excised and fixed for 1 hour in 4% paraformaldehyde. After a brief wash in PBS, corneas were flat-mounted on a slide, mounted with media containing DAPI (VectaShield; Vector Laboratories Inc., Burlingame, CA) and placed at 4°C for 48 hours to allow DAPI penetration. Confocal analysis was performed using a spectral laser scanning confocal microscope (TCS-SP; Leica, Wetzlar, Germany) with the corresponding lasers for DAPI and EGFP. Z-stacks were generated in 0.48-μm step increments, and 3D reconstructions were performed using software (Volocity; Improvision Inc., Lexington, MA). 

In Vivo Inhibition of EGFP Inflammatory Cell Migration in the Corneal Stroma with Anti-MIP2

To recruit EGFP inflammatory cells into the corneal stroma, a 1-mm superficial scratch was made on the corneal surface of each EGFP chimeric mice on the temporal side of the paracentral cornea using a 30-gauge needle and was repeated after 24 hours. Twenty-four hours later, the corneas were imaged for EGFP-positive cells in the area of the scratch using in vivo fluorescence stereomicroscopy. Immediately after imaging, 25 μg MIP-2 neutralizing antibody (R&D Systems, Minneapolis, MN) or 25 μg normal rat IgG (Sigma, St. Louis, MO) in a 2-μL volume was injected intrastromally on the nasal side in the paracentral cornea, directly opposite the scratch injury. Fifteen minutes after injection of MIP-2 neutralizing antibody, 0.5 μg conjugated LPS from E. coli (AlexaFluor 594; Molecular Probes) in a 0.5-μL volume was injected into the same area of the cornea as the MIP-2 or control IgG. Twenty-four hours after the injection of LPS and MIP-2 antibody or control IgG, in vivo fluorescence stereomicroscopic images were captured using identical exposure times and appropriate filters for EGFP-positive cells (green) and the LPS (red) (AlexaFluor 594; Molecular Probes). The area of the cornea directly between the scratch injury and LPS injection was imaged using fluorescence stereomicroscopy, and EGFP density was calculated (Image Pro Plus; Media Cybernetics Inc.). 

Results

In Vivo Visualization of Inflammatory Cells in the Corneas of EGFP Chimeric Mice in Endotoxin-Induced Keratitis

To determine in vivo the recruitment pattern and to track the fate of inflammatory cells during an inflammatory response in the eye, we generated mice that express EGFP in most of their hematopoietic cells (EGFP chimeric mice). EGFP chimeric mice were created by reconstituting the immune system of lethally irradiated C57BL/6 mice with bone marrow–derived cells from EGFP transgenic mice (ACTbEGFP) that express EGFP under the control of chicken β-actin promoter. ^{18 19} Two weeks after bone marrow transplantation, EGFP-positive cells were detected in the bone marrow (46.7%), spleen (78.2%), and blood (39%) of EGFP chimeric mice and none in wild-type C57BL/6 mice (Fig. 1) . Polymorphonuclear cells constituted 17.4% of EGFP-positive cells in the peripheral blood. 

To examine whether the recruitment of EGFP inflammatory cells could be captured in real-time in the cornea, the stroma of EGFP chimeric mice was injected with LPS, PBS (control), or nothing (naive), and the influx of EGFP-positive cells was monitored using in vivo intravital microscopy. Stereoscopic fluorescence in vivo images of anesthetized mouse eyes showed that EGFP-positive cells could be detected in the cornea by 24 hours after injection and that these cells localized to the limbal area (Fig. 2A) . Although EGFP cells were present in the limbus 6 hours after injection, their numbers were too few to document and quantify. No EGFP-positive cells could be detected at earlier time points (data not shown). Mean green fluorescence density analysis of individual corneas at different time points demonstrated that by 48 hours cell recruitment had peaked and, at this time, extended to the paracentral and central cornea. EGFP-positive cells could be seen up to 5 weeks compared with control EGFP chimeric mice that were injected with PBS (Fig. 2B) . These experiments demonstrate that EGFP chimeric mice can be successfully used to track the recruitment and fate of inflammatory cells into the eye in real time. Moreover, they also confirm previous observations and demonstrate clearly that cell migration into the central cornea does not occur until 24 hours after injection of LPS. ⁸  

Characterization of EGFP-Positive Inflammatory Cells in the Corneal Stroma

Because EGFP chimeric mice are generated by injecting whole bone marrow cell preparation from transgenic EGFP mice (ACTbEGFP), the EGFP-positive cells detected in the eye after LPS injection could represent multiple, distinct hematopoietic lineages. To determine the identity of these cells, immunohistochemical staining for EGFP (red), neutrophils (NIMPR-14 green), and macrophages (CD11b green) was performed in 5-μm frozen sections of corneas removed 24 hours after injection (Fig. 3A) . Corneas from EGFP mice injected with LPS showed an increased amount of EGFP-positive infiltrating neutrophils (133) compared with control PBS-injected corneas (83), as determined by staining with the neutrophil-specific stain NIMP-R14 (Fig. 3B)(NIMP-R14: P = 0.006). Given that approximately 50% of the hematopoietic cells in our chimeric mice were EGFP positive, we also detected EGFP-negative neutrophils in the cornea (Fig. 3A) . Not surprisingly, macrophages could also be detected 24 hours after the injection of LPS and could represent a minority of the population of inflammatory cells recruited to the cornea (Fig. 3B) . These data demonstrate that most EGFP-positive hematopoietic cells seen in the corneas of EGFP chimeric mice after LPS injection represent an inflammatory response and confirm that neutrophils migrate first to the corneal stroma during the development of keratitis. 

Infiltration and Migration of Inflammatory Cells through the Corneal Stroma

Real-time in vivo images obtained using stereomicroscopic analysis demonstrate that a significant number of EGFP-positive cells are recruited to the corneal stroma after LPS injection. Moreover, real-time images demonstrate that there was a distinct migration pattern within the corneal stroma as cells migrated through the limbus, paracentral cornea, and central cornea. These observations suggest that the information obtained from the immunohistology of 5-μm sections was incomplete and did not represent all the data from in vivo experiments. 

To circumvent this limitation, we performed confocal analysis of full-thickness whole-mount flat corneas removed 24 hours after the injection of LPS (Fig. 4A) . Images of whole-mount corneas using fluorescence microscopy were used to identify areas of the cornea in which EGFP-positive cells were present for confocal analysis (Fig. 4B) . Stacked confocal images were reconstructed and analyzed in 3D views with the use of software (Volocity; Improvision Inc.). DAPI was used to stain the nuclei of the epithelium and endothelium to delineate these two layers and to help us identify the corneal stroma area between. Moreover, the reconstructed images could be virtually rotated and allowed a comprehensive 3D visualization of the corneal microenvironment. As recently described, we also detected a sporadic distribution of EGFP-positive cells in the corneal stroma of naive mice that have been irradiated ¹⁹ (Fig. 4C) . In contrast, 3D reconstruction analysis of an LPS-injected cornea showed a significant number of EGFP-positive cells migrating through the corneal stroma in a fairly organized fashion (Fig. 4D) . Most inflammatory cells were found to be localized in the anterior stroma and to migrate to areas of the cornea where no EGFP-positive cells had been present. The ability to virtually rotate the 3D reconstructed images of whole-mount corneas not only confirmed observations from the in vivo studies but allowed the visualization of the cornea from different angles and provided a unique and representative image of the interactions of migratory inflammatory cells and the stromal microenvironment. Analysis of these images suggests that the movement of inflammatory cells in the cornea is organized and localized to a specific area of the anterior corneal stromal and seems to follow a path directed by distinct signals. 

Migration of EGFP-Positive Cells Is Influenced by LPS in the Corneal Stroma

Data obtained from the in vivo and ex vivo 3D analysis of whole-mount corneas could be explained by previous studies that show inflammatory cells migrate to the cornea in response to chemotactic signals from the CXC chemokine family induced by LPS and bacterial products. ^{5 8 22} To further investigate in vivo the relationship of cell recruitment, migration, and fate of inflammatory cells with respect to the location of LPS in the corneal stroma, red LPS was injected in the corneal stroma, which had been previously used to track the binding and intracellular processing of LPS in vitro. Using in vivo intravital stereomicroscopy, we found that red LPS was easily detected in the corneal stroma, and this was found to be present at the site of injection and approximately dispersed through half the cornea (Fig. 5A) . Images of the whole corneal mount clearly demonstrate the site of injection, the distribution of red LPS, and the presence of EGFP-positive cells migrating in the stroma (Fig. 5B) . Moreover, EGFP-positive cells were found to be moving away from areas of the corneal limbus that were 180° from the area where most of the LPS accumulated and no LPS was present. As in previous experiments using non-labeled LPS, injecting red LPS into the corneas of EGFP chimeric mice induced the recruitment of EGFP-positive cells 6 hours after injection and followed the same kinetic pattern (data not shown). 

To determine the relationship between LPS and the migration pattern of EGFP-positive cells coming into the LPS depot, we analyzed by confocal microscopy two areas with different densities of LPS. We defined zone I as next to the site of injection (with a significant amount of LPS) and zone II as away from the site of injection, where LPS was highly diluted (Fig. 5B) . Three-dimensional reconstruction analysis showed that EGFP-positive cells could be found within the red particles of LPS at 24 and 48 hours; however, no colocalization was observed where LPS was present at dense (zone I) or limited (Zone II) concentrations (Figs. 5C and 5D) . Moreover, images obtained in zone II showed EGFP-positive cells in the anterior portion of the corneal stroma where LPS was present at low concentrations and where LPS was present at higher concentrations. These observations are consistent with a cell migration pattern of inflammatory cells in the corneal stroma. 

Notably, EFPG-positive cells did not interact or bind to LPS particles in the corneal stroma. This was demonstrated using 3D tilted views of areas to visualize that inflammatory cells were moving along the path where the LPS was in zone II, toward (but not localizing to) the site of injection where a dense area of LPS was present (Figs. 6A 6B 6C) . This suggested and confirmed previous observations that the recruitment of inflammatory cells is determined by other signals or soluble factors induced by the LPS in the corneal stroma. ^{5 8}  

In Vivo Blockade of MIP-2 and Inflammatory Cell Migration in the Corneal Stroma

Three-dimensional confocal analysis demonstrated that the migration of EGFP inflammatory cells (mostly neutrophils) in the cornea seemed to be guided by other signals induced by LPS. MIP-2 is a chemokine of the CXC family, which is a murine functional homologue of IL-8 important in the recruitment of neutrophils in inflammation. ^{23 24} MIP-2 has been shown to have an important role in the recruitment and extravasation of neutrophils in keratitis. ^{5 8} To determine whether MIP-2 also is responsible for the migration of EGFP-positive cells through the corneal stroma after extravasation, we first recruited EGFP-positive cells into the stroma by performing a series of corneal scratches localized to a specific area of the paracentral cornea. Once these cells were localized, we neutralized MIP-2 in the corneal stroma using anti-MIP2 antibody injected into the stromal space before the injection of red LPS in the opposite site of the cornea. ^{5 8 22} In vivo analysis at 12 and 24 hours after LPS injection of EGFP chimeric mice treated with anti–MIP-2 demonstrated that the migration of EGFP-positive cells through the corneal stroma was significantly inhibited, but not totally eliminated, by 47% (P < 0.0001) compared with animals given normal rat IgG or with PBS-injected control mice (Fig. 7) . These results suggest that the production of MIP-2 in the corneal stroma not only is responsible for inflammatory cell recruitment to the limbal vessels but may be an important signal that guides the migration of these cells through the stroma in response to the accumulation of LPS. It is important to note that in this group of experiments extravasation was also affected by anti-MIP2 treatment. Therefore, the effect of extravasation on the magnitude of cell migration cannot be completely ruled out, and the role of MIP-2 in this process may still be the major determinant of how cells migrate through the corneal stroma. 

Discussion

The in vivo study of immune responses “in situ” is not simple. The recruitment, migration, and cellular interactions of inflammatory cells with the extracellular matrix are difficult to mimic in tissue culture systems. The eye provides a unique window into the body through which the visualization of immunologic events can be tracked in real time. ¹⁵ The experiments presented in this work represent an innovative approach that combines the use of in vivo and ex vivo imaging techniques with EGFP-positive bone marrow–derived cells to determine the recruitment and fate of inflammatory cells during an immunologic response in the cornea. 

The generation and use of EGFP chimeric mice in these experiments is an approach that has made in vivo real-time imaging a very sensitive method of tracking and quantifying inflammatory cell recruitment into sites of inflammation. We have taken full advantage of the accessibility and transparency of the cornea to visualize and characterize in a unique way the development of an inflammatory response induced by endotoxin. The infiltration of the corneal stroma by inflammatory cells results in the disruption of highly organized collagen fibril matrix deposition, dysfunction, and loss of clarity that can lead to significant visual impairment. ²⁵ Earlier studies have demonstrated how the presence of neutrophils in the cornea, in response to LPS, can disrupt its normal architecture by releasing proteolytic enzymes from their granules. ^{26 27} Our findings on the presence and migration of neutrophils through distinct lamellar planes of the corneal stroma are consistent with this report and raise the possibility that proteolytic enzymes produced by neutrophils are important in making the extracellular matrix of the corneal stroma accessible for cellular migration. Moreover, our observation that this process is also dependent on the chemokine MIP-2 is novel. 

The unique views produced by this technique provided a different anatomic perspective with regard to the pathway and location of inflammatory cells during their migration to the cornea. Previous studies have demonstrated how neutrophils in infectious or endotoxin-induced keratitis are primarily recruited to the cornea through limbal and conjunctival vessels. ^{4 28} Data generated from our studies, specifically the experiments in Figure 7 , help us determine subsequent events. The migration of EGFP-positive cells within the cornea follows a predominant path in the anterior stroma. Cells migrate in an organized fashion toward the deposit of LPS in the corneal stroma. Although, this could be attributed to hydrodissection of collagen lamellae induced by the intrastromal injection of LPS, the images obtained after the injection of red-labeled LPS showed that these cells traveled toward LPS and suggested that a gradient of chemotactic signals develops in response to endotoxin. We cannot rule out the possibility that this specific pattern of migration could be a result of the breakdown of the collagen matrix by factors produced by infiltrating neutrophils. The process of leukocyte recruitment and extravasation from the vessel into the tissue is complex and depends on the interaction of inflammatory cells and integrins and other adhesion molecules expressed in the vascular endothelium. ^{28 29} Our model suggests that, indeed, cells then migrate through the tissue, after another cascade of complex steps mostly mediated by chemokines; how adhesion molecules interact could be dissected with this in vivo model. 

The fact that EGFP-positive inflammatory cells do not interact with LPS in the corneal stroma suggests that these chemotactic signals may be coming from another source or another cell that interacts with LPS. The rotation of 3D-reconstructed views of inflamed corneas demonstrates that EGFP cells travel over and through areas in which LPS is deposited. In vitro studies have shown that corneal stromal keratocytes produce chemokines and cytokines in response to proinflammatory signals. ^{30 31} Moreover, mice with reduced LPS responses resulting from a deficiency in Toll-like receptor-4 (TLR4) signaling have also been shown to have a reduced corneal inflammatory response to LPS. ⁸ Therefore, it is tempting to postulate and to predict that stromal keratocytes interact with the injected LPS to produce chemotactic factors that create a chemotactic gradient in the corneal tissue that is responsible for guiding EGFP inflammatory cells toward the site of injection, as shown by our 3D confocal images. It is important to realize that infiltrating neutrophils are also a source of chemokines; how they affect the migration pattern of other inflammatory cells remains to be explored. 

Based on the observations published by other groups, we investigated the role of MIP-2 in the migration of EGFP inflammatory cells through the corneal stroma. ^{5 8 22 32} The reduced inhibition of EGFP-positive cell migration in the corneal stroma by the injection of anti–MIP-2 neutralizing antibodies implies that this chemokine could also be responsible for the migration of neutrophils (Fig. 7) . The fact that there was not a complete inhibition of cellular migration into the cornea means that other important chemokines, such as MIP-1α, are also involved in this process. ³³ Moreover, it is important to realize two limitations of the study are that differentiation between extravasation and migration is not totally possible and that the effects of MIP-2 in extravasation could also be the major determinant of cell migration in the corneal stroma. 

In conclusion, the use of EGFP chimeric mice combined with in vivo and ex vivo imaging technology can be effectively used to track the recruitment and fate of bone marrow–derived inflammatory cells during an inflammatory response to LPS in the eye. New and unique biologic observations can be generated from in vivo experiments using this model. We have demonstrated that the migration of inflammatory cells through the corneal stroma is distinct and seems to follow an established chemotactic gradient mainly by the chemokine MIP-2, which is also involved in the recruitment phase of this process. We believe that the novel methods of visualizing immune responses in the eye described in this work will be useful in the study of immunologic reactions in autoimmunity and allotransplantation. 

 Supported by National Institutes of Health Grant K08EY014912–01 (VLP), NIH Infrastructure Grant R24EY15638, and by the Knights Templar Foundation (VLP).

 Submitted for publication June 23, 2004; revised March 14 and July 31, 2005; accepted October 21, 2005.

 Disclosure: E.C. Carlson, None; J. Drazba, None; X. Yang, None; V.L. Perez, 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.

 

Reconstitution of spleen, bone marrow, and peripheral blood lymphocytes of naive and irradiated C57BL/6 mice with bone marrow cells isolated from C57BL/6-TgN(ACTbEGFP)1Osb mice. Cells from the spleen, bone marrow, and peripheral blood were harvested and analyzed by flow cytometry to measure the expression of EGFP in naive C57BL/6 and EGFP chimeric mice. The spleen, bone marrow, and total blood show reconstitution percentages of 78.2%, 46.7%, and 39.1%, respectively, in the EGFP chimeric mice. Lymphocyte and polymorphonuclear populations expressing EGFP in total blood preparations are also shown.

Figure 2.

 

In vivo fluorescence stereomicrographs (32× magnification; 1.5-second exposure) and quantification of the recruitment of EGFP-positive inflammatory cells in corneas of EGFP chimeric mice injected with LPS or PBS. Recruitment of EGFP-positive cells to the cornea was monitored in vivo at multiple time points (6 hours to 5 weeks) after injection of PBS or LPS by fluorescence microscopy. (A) Accumulation and resolution of EGFP-positive inflammatory cells in the cornea from 6 hours to 5 weeks after injection. (B) Mean green fluorescence density values demonstrate the increase in keratitis, which begins to resolve between 5 and 7 days.

Figure 3.

 

Detection and quantification of EGFP-positive neutrophils and macrophages in inflamed corneas of EGFP chimeric mice. Corneas from PBS control and LPS injected EGFP chimeric mice were harvested at 24 hours for immunohistochemistry. (A) Sections were stained with anti-EGFP (red) and anti-neutrophil antibody (NIMP-R14; green) and then merged. Images demonstrate the presence of EGFP-positive neutrophils infiltrating the corneal stroma. Image with DAPI stain and no primary antibody demonstrates that green EGFP from chimeric cells cannot be readily detected in these sections. (B) Central cornea sections were analyzed, and the number of neutrophils (NIMP-R14) and macrophages (F4/80) were counted. Graphs demonstrate that at 24 hours most EGFP inflammatory cells were neutrophils (P = 0.006); however, a population of macrophages was also present in the cornea. Results represent the mean ± SD of three central corneal sections per group.

Figure 4.

 

Whole-mount corneas and 3D reconstruction analysis of confocal images of EGFP chimeric mouse corneas 24 hours after injection of LPS. (A) Flat-mounted cornea. (B) Fluorescence image of the flat-mounted cornea of an EGFP chimeric mouse 24 hours after injection with LPS. EGFP-positive cells can be seen in different areas of the corneal stroma. (C) Three-dimensional reconstructed image of a cornea from a noninjected EGFP chimeric mouse shows a sporadic distribution of EGFP-positive cells in the corneal stroma in contrast to an LPS injected cornea, (D) which shows the presence of EGFP-positive inflammatory cells migrating through an anterior stromal plane in the cornea.

Figure 5.

 

Whole-mount corneas and 3D reconstruction analysis of confocal images of EGFP chimeric mouse corneas 24 hours after injection of AlexaFluor 594 (red)–conjugated LPS. (A) In vivo stereomicroscopic fluorescence image (32×) demonstrates the location of LPS in the corneal stroma after injection. (B) Fluorescence image of the flat-mounted cornea of an EGFP chimeric mouse 24 hours after injection with red LPS. From this image, an area with dense deposition (zone I) and minimal deposition (zone II) of LPS was determined for 3D confocal analysis (C, D). Side view pictures (400×) from the 3D reconstruction analysis demonstrate that EGFP-positive cells migrate through the particles of red LPS in zone I and are moving toward the red LPS in zone II in an anterior stromal plane, as previously described.

Figure 6.

 

Three-dimensional reconstructed DAPI-stained (blue) corneal whole-mounts from EGFP chimeric mice injected with AlexaFluor 594–conjugated LPS (red), as viewed from the top (A), –30° y-axis rotation (B), and a tilted view from the side (C) demonstrate the migration pattern and distribution of EGFP-positive cells through the corneal stroma anterior to the deposit of LPS.

Figure 7.

 

In vivo inhibition of EGFP-positive inflammatory cell corneal stroma migration by anti–MIP-2 treatment of LPS-induced keratitis. EGFP chimeric mice received intrastromal injections of anti–MIP-2 neutralizing antibody or immunoglobulin control immediately before intrastromal injection of LPS. The graph demonstrates the inhibition of EGFP-positive cell migration through the corneal stroma in the area of anti–MIP-2 treatment, as determined by mean green fluorescence density from in vivo stereomicroscopic images obtained 12 and 24 hours after injection. Results are mean green fluorescence density values ± SD of 3 mice per group. Analyses (t test) demonstrate a statistical (P < 0.001) difference between scratch and LPS-injected corneas and, more important, between anti–MIP-2 and IgG control (P < 0.0001) groups.