Accelerated Wound Healing of Alkali-Burned Corneas in MRL Mice Is Associated with a Reduced Inflammatory Signature

Abstract

purpose. The present study was conducted to investigate healing of alkali-burned corneas in MRL/MpJ (MRL) mice.

methods. Gross, clinical, and histologic criteria were used to compare healing of alkali-burned corneas in MRL and control C57BL/6J (B6) mice. Effects of neutrophil depletion of B6 mice and allogeneic reconstitution of B6 mice with MRL bone marrow on wound healing were evaluated. Gene expression patterns in normal and wounded corneas were surveyed with array-based quantitative real-time RT-PCR (AQPCR).

results. MRL mice showed accelerated reepithelialization and decreased corneal opacity compared with B6 mice after alkali wounding. Marked inflammatory cell infiltration and fibrosis were evident in the corneas and anterior chambers of B6 mice. MRL mice showed less severe lesions, except for stromal edema. Rapid reepithelialization and reduced keratitis/iritis were also observed in neutrophil-depleted B6 mice, but not in B6 mice reconstituted with MRL bone marrow. AQPCR showed transcriptional changes of fewer genes associated with inflammation and corneal tissue homeostasis in alkali-burned corneas from MRL mice. Increased expression of an anti-inflammatory gene, Socs1, and a gene associated with healing, Mmp1a, were evident in MRL corneas.

conclusions. Alkali-burned corneas heal faster and more completely in MRL mice than in B6 mice, by means of rapid reepithelialization, reduced inflammation, and reduced fibrosis. Reduced inflammation, including decreased neutrophil infiltrates and the lack of a robust proinflammatory gene expression signature correlates with the rapid healing. However, the rapid-healing phenotype is not intrinsic to MRL hematopoietic progenitor cells.

Corneal wound healing is a complex process requiring integrated function of multiple tissues, cell lineages, growth factors, and cytokines. ^{1 2 3 4} Rapid reepithelialization is critical for successful corneal healing. Impaired reepithelialization increases the risk of infection and is associated with heightened inflammation and insufficient stromal remodeling, resulting in loss of transparency of corneal tissues. Even the small scars in the cornea that occur during the normal wound-healing processes, can lead to severe loss of corneal transparency and consequent visual impairment. The most common response of mammalian tissue to wounding is a repair process characterized by inflammation accompanied by extracellular matrix remodeling. This process leads to formation of scar tissue at the lesion site. ⁵ Despite the need for improved treatment of corneal injury, the mechanisms of corneal wound healing are not fully understood. Thus, the identification of animal models showing rapid corneal wound healing and determination of mechanisms involved in the healing process may lead to the identification of therapeutic targets and the development of effective therapy. 

The MRL/MpJ (MRL) inbred mouse strain was originally derived through a series of crosses involving strains LG/J, AKR/J, C3H/Di, and C57BL/6. ⁶ Previous studies have shown that adult MRL mice display complete closure of full-thickness ear punch holes. The healing occurs without scar tissue formation and with reconstitution of normal architecture that includes hair follicles, sebaceous glands, and cartilage. In contrast, the ear punch holes made in other mouse strains never close, and surrounding tissues contain abundant scar tissue. ⁷ The regenerative healing of MRL mice has also been described after injury to the heart. ⁸ These findings suggest the possibility of MRL mice as a model of improved corneal wound healing. 

Alkali burn, characterized by a high degree of penetration of the cornea, typically causes severe injury of corneal tissues. It often results in a marked inflammatory cell infiltration, extensive formation of scar tissue, recurrent epithelium erosion, ulceration, stromal edema, and neovascularization. These processes lead to decreased corneal transparency. ⁹ Alkali-burned corneas seldom heal properly and show reduced corneal transparency. Prevention of scar tissue is an important clinical consideration after corneal injury. ¹⁰ However, the precise mechanisms of alkali injury and the wound-healing response remain poorly understood. 

In the present study, we investigated healing of alkali-burned corneas in MRL mice. These mice showed accelerated corneal wound healing compared with C57BL/6J control animals. The rapid corneal wound healing in MRL mice was accompanied by lowered inflammatory responses and highly activated reepithelialization. 

Materials and Methods

Animals

Male MRL/MpJ (MRL) and C57BL/6J (B6) mice at 6 to 12 weeks of age were obtained from The Jackson Laboratory (Bar Harbor, ME). Animals were handled according to the guidelines in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The Institutional Animal Care and Use Committee (IACUC) of The Jackson Laboratory approved all animal protocols in this study. 

Alkali Burn

Alkali burning of corneas was performed with mice under anesthesia with systemic ketamine and xylazine followed by topical tetracaine. NaOH (4 μL of a 0.2-M solution per eye) was instilled on the entire surface of both corneas. The eyes were irrigated 30 seconds later with 10 mL of artificial tears (Earle’s balanced salt solution with extra KCl, to a final concentration of 1.0 g/L, to make the electrolyte composition similar to natural tears). ¹¹ Immediately after irrigation, the eyes were stained with 1 mg/mL fluorescein prepared from fluorescein sodium ophthalmic strips (Fluorets; Chauvin, Aubenas, France). Fluorescein staining of the entire cornea confirmed that all corneal regions were wounded. Erythromycin ointment was then applied to prevent corneal surfaces from drying during anesthesia. 

Clinical Observations and Histopathology

Corneal opacity was classified under a dissection microscope as follows: 0, no opacity; 1, less than one third of the corneal surface is clouded; 2, less than two thirds of the corneal surface is clouded; 3, more than two thirds of the corneal surface is clouded; and 4, almost all the corneal surface is clouded, and the opacity prevents visualization of the pupil margins. Corneal reepithelialization was evaluated under a dissection microscope by assessing the degree of corneal staining with fluorescein, classified as open (not fully reepithelialized) or closed (fully reepithelialized). For histologic examination, mice were euthanatized by cervical dislocation, and the eyes were fixed in 10% buffered formalin, embedded in paraffin, and processed for light microscopy. Sections cut at 4 μm were stained with hematoxylin-eosin. 

In Vivo Neutrophil Depletion

B6 mice received intraperitoneal injections with either PBS or 150 μg/mouse of anti-mouse Gr-1 monoclonal antibody (RB6-8C5, obtained from unpurified ascites and sterile filtered) at 1 day before and 1, 3, and 5 days after alkali burn. Administration of this antibody causes specific reduction in the number of circulating neutrophils. ¹² To confirm the effects of antibody treatment, peripheral blood (50 μL) was collected from the tail at 1 day before and 3 and 7 days after alkali burn. Total WBC count and WBC differential counts were performed (ADVIA 120 Hematology System; Bayer Corp., Tarrytown, NY) as previously described. ¹³  

Bone Marrow Transplantation

Bone marrow cells of B6 or MRL mice were recovered aseptically from femurs and tibias after they were crushed with a mortar and pestle. The bone marrow cells were resuspended in RPMI medium containing 10% FBS. The recipient B6 mice were exposed to 900 cGy of whole-body radiation from a Cs¹³⁷ source. All recipient mice received a single intravenous injection of 25 × 10⁶ nucleated donor bone marrow cells via the lateral tail vein. For B6 mice treated with allogeneic (MRL) bone marrow cells, 0.5 mg/mouse of anti-mouse CD-154 (CD40 ligand) monoclonal antibody (MR-1, obtained from unpurified ascites and sterile filtered) was administered intraperitoneally immediately and 3 and 14 days after transplantation. The transient blockade of the CD40–CD-154 costimulatory pathway using this antibody to treat lethally irradiated recipients has been reported to establish a robust allogeneic hematopoietic chimerism without subsequent graft-versus-host disease. ¹⁴ Mice injected with syngeneic (B6) bone marrow cells received PBS in the same manner as the antibody treatment. Five weeks after transplantation, blood (approximately 100 μL) was collected from tails of the bone marrow recipients and from the untreated B6 and MRL mice. The percentage of peripheral blood mononuclear cells (PBMCs) expressing B6 (H2-K^b) and MRL (H2-K^k) major histocompatibility complex class I antigens was determined by dual labeling with anti-H2-K^b (28-13-3s) and anti-H2-K^k (36.7.5) monoclonal antibodies. The stained cells were analyzed by flow cytometry (FACSCalibur; BD Biosciences, San Jose, CA), as previously described. ¹³  

Gene Expression Analysis

The corneas were collected from sham surgery or alkali-burned mice of both strains 7 days after the treatments, and they were immediately placed into an RNA stabilization reagent (RNAlater; Ambion Inc., Austin, TX). RNA from each mouse (two corneas per mouse) was isolated (RNeasy Micro Kit; Qiagen Inc., Valencia, CA) according to the manufacturer’s protocol, and yields ranged from 40 to 180 ng. Gene expression patterns were investigated with an array-based quantitative real-time RT-PCR (AQPCR) according to the method of Akilesh et al. ¹⁵ The array-based QPCR consists of 384 PCR amplicons that survey, at the transcriptional level, genes associated with a broad spectrum of immunologic processes. A listing of these 384 genes is available online at http://www.jax.org/staff/roopenian/labsite/gene_expression.html. 

Statistical Analysis

All data, except for those from the AQPCR, were analyzed by two-sided, two-group comparisons on computer (Prism, ver 4.0; GraphPad Software, San Diego, CA). Continuous, ordered categorical, and binary data were analyzed by unpaired t-test, Mann-Whitney test, and the Fisher exact test, respectively. P < 0.05 was considered statistically significant. For data from the array-based QPCR, cycle thresholds generated by a rapid PCR sequence-detection system (Prism 7900HT; ABI) were analyzed by the global pattern recognition (GPR) algorithm ¹⁵ (version 2.0, implemented as a macro in Excel; Microsoft, Redmond, WA). GPR parameters for accepting data were a cycle cutoff of ≥37.5, a threshold at P ≤ 0.05, and a GPR score of ≥0.4. 

Results

Accelerated Corneal Wound Healing in MRL Mice

The levels of healing of alkali-burned corneas were compared between B6 and MRL mice. Table 1shows the rate of corneal reepithelialization in both strains. MRL mice showed accelerated rates of reepithelialization. Statistically significant differences in the number of healed (fully reepithelialized) corneas were evident on days 3 through 7. As indicated in Table 1 , small corneal ulcerations, as judged by fluorescein staining, recurred after initial healing in 4 of 28 B6 eyes and 9 of 30 MRL eyes. For all eyes examined, these small ulcerations healed within 2 days (data not shown). Figure 1shows the time course of corneal opacity. Severe corneal opacity occurred by day 4 in both strains. Although, the corneal opacity in MRL mice gradually decreased from days 7 through 28, the corneal opacity in B6 mice remained severe at day 28. There were statistically significant differences in mean opacity scores on days 7, 14, and 28. As shown in Figure 2 , the gross appearance of eyes in MRL mice on day 28 (Fig. 2J)was similar to that of healthy, nontreated mice (not shown), whereas severe opacity and neovascularization were still present in B6 mice (Fig. 2E) , indicating corneal scarring. Figure 2shows the results of histologic examination. Severe keratitis and iritis were present on days 7 and 14 in B6 mice (Figs. 2B 2C) . Corneal lesions consisted of stromal edema, keratocyte loss, ulceration, inflammatory cell infiltration, and neovascularization. Other lesions included hyphema; inflammatory cell infiltration, predominately neutrophils, with a lesser number of macrophages; fibrosis in the anterior chamber; and anterior lens epithelial hyperplasia. Moderate inflammation was still present on day 28 (Fig. 2D) . In contrast, MRL mice (Figs. 2G 2H 2I 2J)had mild inflammation in the corneas and anterior chamber. However, on days 7 and 14 in MRL corneas, the stromal edema was as severe as that observed in B6 mice. Only mild stromal edema remained in MRL corneas on day 28. Histologic examinations were also conducted on corneas sampled on days 1 through 4, when the significant changes in reepithelialization were seen in MRL mice. The results are shown in Figure 3 . Severe keratocyte loss was observed in both strains. There was massive infiltration of inflammatory cells in the corneal stroma and anterior chamber in B6 mice beginning on day 1. Inflammation was most severe on day 2 for the cornea and day 4 for the anterior chamber. In contrast to B6 mice, MRL mice had minimal inflammation throughout the observation period. 

Effects of Depletion of Circulating Neutrophils on Corneal Wound Healing after Alkali Burn in B6 Mice

Because minimal inflammation was noted in MRL (fast healer) but not in B6 (slow healer) mice, we speculated that inflammatory cells, especially neutrophils, retard reepithelialization and matrix remodeling. We tested this hypothesis by treating B6 mice with monoclonal anti-Gr-1 antibody. As shown in Figure 4 , the anti-Gr-1 treatment caused an 82% reduction in the number of circulating neutrophils on day 3. A 34% reduction in neutrophil numbers was observed on day 7, although it was not statistically significant (P = 0.0513). There was a 34% increase in peripheral lymphocyte numbers on day 7 in the anti-Gr-1-treated mice, which contributed to a 25% increase in the number of total leukocytes. As shown in Table 2 , the anti-Gr-1-treated mice showed faster corneal reepithelialization than the PBS-treated mice. This difference was statistically significant on days 4 and 5. The timing of accelerated corneal reepithelialization correlated with the time of greatest granulocyte depletion in the anti-Gr-1-treated mice. Histologic examination of eyes sampled on day 7 (Fig. 5)showed decreased inflammatory cell infiltration into the cornea and anterior chamber, as well as reduced fibrosis and hyphema in the anterior chamber of the anti-Gr-1-treated mice. Therefore, depletion of circulating neutrophils resulted in reduced inflammation and accelerated healing of alkali-burned corneas in B6 mice. 

Effects of Transplantation of MRL Bone Marrow on Corneal Wound Healing in B6 Mice

We next examined whether allogeneic reconstitution of the hematopoietic system of B6 mice with MRL bone marrow would result in accelerated healing of corneas after alkali burning. B6 bone marrow chimeras were established with MRL allogeneic bone marrow or B6 syngeneic bone marrow. Costimulation blockade was used to establish the MRL bone marrow chimeras. Figure 6shows representative flow cytometry of PBMCs stained for H2-K^k (MRL origin) and H2-K^b (B6 origin). The relative percentage of MRL-,origin cells in recipient B6 mice was >96% in all MRL bone marrow chimeras. As shown in Table 3 , there were no differences in rate of corneal reepithelialization between B6 and MRL chimeras. There was an overall delay in the corneal reepithelialization of both bone marrow chimeras compared with reepithelialization in nonirradiated animals (Table 1) . Histologic examination of eyes sampled on day 7 (Fig. 7)showed that the MRL chimeras had inflammatory changes as severe as those in the B6 chimeras. Thus, hematopoietic engraftment with MRL bone marrow cells alone did not result in either rapid reepithelialization or reduced inflammation, as was seen in MRL mice. 

Gene Expression Analysis of Alkali-Burned Corneas

To gain an understanding of the transcriptional basis of corneal injury, we surveyed gene expression patterns of sham or partially healed corneas of B6 and MRL mice. Of the 384 array genes examined, only two, H28 and Flt1, were differentially expressed when sham-surgery corneas of both strains were compared (Table 4A) . H28 encodes a type I interferon-responsive gene expressed by MRL mice, whereas B6 mice do not express this gene. ¹⁶ MRL mice had higher expression of Flt1, vascular endothelial growth factor (VEGF) receptor 1, presumably also reflecting a transcriptional polymorphism in this gene. Nineteen genes were quantitatively different in expression in B6 alkali-burned corneas when compared with sham-surgery B6 corneas (Table 4B) . Downregulation of Satb1, a nuclear protein, is likely to be a reflection of the death of B6 corneal or infiltrating cells, because a decreased protein level has been reported in cells undergoing apoptosis or necrosis. ¹⁷ Many of the genes differentially expressed in the alkali-B6 versus sham-surgery B6 corneas were related to inflammation. These genes include: Cd48, Spp1, Tyrobp, Tnfsf13b, Icam1, Inpp5d, Cd14, Sfpi1, Il1r2, Pfkp, Fcgr3, and Ly6a, all of which were upregulated, except for Tnfsf13b. Flt1 and Enpep genes are involved in angiogenesis. ^{18 19} Flt1 was upregulated and Enpep was downregulated. The Pax6, Bcl2, Hspa2, and Mmp13 genes are associated with homeostasis or wound healing of corneal tissue. ^{20 21 22 23} In contrast to the many genes upregulated in alkali-burned B6 corneas, alkali-burned MRL corneas showed expression changes in fewer genes compared with sham-MRL corneas (Table 4C) . Enpep, Ly6a, Icam1, and Tyrobp showed expression alterations similar to those observed in the above B6 mice. The only changes unique to the MRL response to alkali injury were the upregulation of B2m and Ccr7 and a downregulation of H28. A direct comparison between alkali-burned MRL and B6 corneas also revealed expression changes in seven genes that distinguish the transcriptional status of wounded MRL and B6 corneas (Table 4D) . Alkali-burned MRL corneas showed higher expression of Mmp1a, a gene associated with corneal wound healing, ²³ as well as Tnfrsf13c, Ccl11, Ccl28, and Socs1. Lowered expression of the stimulatory NK receptor GP49a and the myeloid–lymphoid transcript Sfpi1 were also observed, potentially reflecting fewer inflammatory cells in the alkali-burned MRL corneas. 

Discussion

In the present study, MRL mice showed accelerated corneal wound healing compared with B6 control mice. After alkali burn, the accelerated corneal wound healing in MRL mice was associated with rapid reepithelialization, reduced inflammation, and reduced fibrosis. In the present study, alkali was placed on the entire corneal surface providing access to all external ocular tissues. Fluorescein staining alone does not prove complete loss of corneal epithelial cells. Although the regenerating epithelial cells in MRL and B6 mice do not morphologically differ from intact corneal epithelial cells, further investigation is needed to clarify the origin of regenerating cells in this alkali-burn model. Some mice of both strains showed recurrent small corneal ulcerations, a well-established clinical finding in alkali-injured patients. ⁹ Rapid healing MRL mice showed a larger incidence of recurrence compared with B6 mice. The increased incidence of recurrent small corneal ulcerations in MRL mice may be associated with increased corneal epithelium at risk. MRL mice can heal full-thickness ear holes in a regenerative manner and without scar formation. Other manifestations of regenerative healing in mammals occur in embryonic and early fetal life. In all cases, rapid reepithelialization at the wound site is a commonly observed feature. ^{24 25 26} Minimal inflammation associated with regenerative healing has also been demonstrated in mammalian embryos and fetuses. ^{25 27} However, the scarless healing in the mouse fetus occurs until embryonic day 16, after which there is a transition to the scarring repair seen in adult skin. That time in mouse ontogeny also coincides with appearance of inflammatory responses to wounds. ²⁸ All evidence taken together shows that it is very likely that reduced inflammation plays an important role in the accelerated corneal wound healing we observed in the corneas of MRL mice. 

In general, the neutrophil is the predominate cell type in the acute phase of inflammation, soon followed by a second wave of monocyte infiltration. In the case of rat alkali-injured corneas, neutrophils are reported to be present as early as 6 hours after injury. ²⁹ Neutrophils and their lysate significantly slow rat corneal epithelial wound healing in vitro, ³⁰ and selective neutrophil suppression by anti-neutrophil antibody also reduces the development of corneal ulcerations in a guinea pig model of alkali burn. ³¹ Accelerated wound skin closure has also been reported in neutrophil-depleted mice. ³² Therefore, we focused on neutrophils, and the effect of their depletion was examined. As expected, the neutrophil-depleted B6 mice showed faster corneal reepithelialization. Although insufficient depletion was achieved on day 7, histologic examination still demonstrated acceleration of corneal healing in the neutrophil-depleted mice. These results suggest that infiltrating neutrophils, at least in part, contribute to a delay in wound healing after alkali injury. The reepithelialization in neutrophil-depleted B6 mice (Table 2)was still delayed when compared with that of MRL mice (Table 1) . One explanation for the partial improvement in wound healing was incomplete neutrophil depletion. It is also possible that inflammatory mediators from lymphoid cell populations in B6 mice delay wound healing. In addition, decreased regenerative activity of epithelial cells or decreased support by stromal cells in B6 corneas may contribute to slow healing. 

What cells or tissues are responsible for the reduced inflammation in MRL mice? It has been reported that MRL mice show decreased neutrophil accumulation in the lung in response to LPS challenge in an experiment with bone marrow chimeras. ³³ This reduction in neutrophil accumulation was dependent on bone marrow cells’ being of MRL origin. Therefore, we examined whether MRL bone marrow cells transplanted into B6 mice would mimic the accelerated corneal wound healing. Contrary to our expectation, B6 mice engrafted with MRL bone marrow cells were not able to facilitate rapid reepithelialization and reduced inflammation. Therefore, factors intrinsic to the hematopoietic microenvironment or host corneal cells rather than hematopoietic cells are suggested to be responsible for manifesting the accelerated healing. Similar findings have been reported for MRL skin wound healing, ³³ but the reverse has been reported for MRL accelerated heart regenerative healing. ³⁴ These discrepancies suggest that the mechanisms of accelerated/regenerative healing in MRL mice is tissue-type dependent. 

Owing to the sensitivity of the array-based QPCR procedure, we were able to develop a robust transcription profile of gene expression in corneas from individual mice. The sham-surgery corneas of B6 and MRL mice revealed minimal natural expression polymorphisms in the steady state. Moreover, when compared on a sham versus alkali-treated basis, Enpep, Ly6a, Icam1, and Tyrobp showed expression alterations in both MRL and B6 mice. These changes can thus be considered to be generic signatures of corneal injury and repair. It was quite clear that MRL mice showed few additional expression changes in response to alkali injury. In stark contrast, B6 mice showed unique changes in the expression of several genes considered as markers for inflammatory cells, including CD14, CD48, and Fcgr3, along with IL1r2. Corneas from B6 mice also showed strong expression of Spp1 (osteopontin), whereas MRL mice showed no upregulation of this gene. Osteopontin is a chemoattractant for inflammatory cells and is synthesized by a variety of cells under pathologic conditions. ³⁵ Osteopontin is also one of the matricellular proteins that is present at the wound site, and it facilitates the earliest aspects of leukocyte infiltration. ³⁶ Spp1 is expressed during fibrotic responses, and Spp1 knockout mice show reduced interstitial fibrosis. ³⁷ There was a downregulation of Tnfsf13b, a cytokine responsible for B-cell survival (BAFF), ³⁸ in B6, but not MRL, mice. Instead, a much higher expression of Tnfrsf13c, a BAFF receptor, was noted in alkali-wounded corneas from MRL mice compared with wounded corneas from B6 mice. Alkali-wounded MRL corneas also had higher expression of Ccl11 (chemokine for eosinophils) and Ccl28 (chemokine for lymphocytes) than did B6 corneas. The concept that the gene expression changes in MRL mice are associated with anti-inflammatory and antifibrotic effects is further supported by the higher expression of suppressor of cytokine signaling-1 (Socs1); SOCS1 suppresses lipopolysaccharide (LPS) signaling in macrophages. Socs1-deficient mice are more sensitive to LPS shock than are their wild-type littermates, and macrophages from these mice produce increased amounts of proinflammatory cytokines. ³⁹ SOCS1-deficient dendritic cells not only induce B-cell proliferation but also can trigger allogeneic T cell expansion. ³⁹ Socs1 ^{− / −} Ifng ^{+/ −} mice have corneal inflammatory infiltration and ulceration ⁴⁰ and Socs1 ^{+/ −} mice have more severe liver fibrosis than do wild-type mice in a liver fibrosis model involving dimethylnitrosamine. ⁴¹ It is therefore possible that at least part of rapid wound healing of MRL corneas is caused by the abundance of Socs1 expression. 

There was an upregulation of Flt1, a receptor for VEGF, an endothelial cell-specific angiogenic and permeability-inducing factor, in B6 mice. Higher expression of VEGF and Flt1 is reported in inflamed and vascularized human corneas. ¹⁸ In contrast, Enpep (glutamyl aminopeptidase) was downregulated in both B6 and MRL mice. Enpep is a proangiogenic factor and Enpep-null mice exhibit an impaired angiogenic response to hypoxia or growth factors. ¹⁹ The downregulation of Enpep may be related to pathogenesis of alkali injury. 

There were changes in gene expression associated with homeostasis or wound healing of corneal tissue in B6 mice. These genes include: Pax6, Bcl2, Hspa2, and Mmp13. MRL corneas did not show such changes. In contrast, MRL mice showed increased expression of Mmp1a. Matrix metalloproteinases (MMPs) are a family of proteolytic enzymes that function to maintain and remodel tissue architecture. MMPs are also implicated in the pathogenesis of ocular surface diseases, especially corneal ulcers. ²³ Both Mmp1a and Mmp13 are upregulated in human corneal epithelial cells (HCECs) in response to inflammatory cytokines, IL-1β, and TNF-α. ⁴² MMP-1 is found in association with HCECs at migratory leading edges of scratch-wound closure. Neutralization of MMP-1 significantly decreases scratch-wound closure in ex-vivo wound-healing experiments. ⁴³ MMP-13 is also reported to be expressed in the epithelial cells at the leading edge of wounded rat corneas and is considered to play a role in reepithelialization after corneal wounding. ⁴⁴ Therefore, it is likely that the increased expression of at least Mmp1a contributes to the accelerated reepithelialization in MRL mice. The gene expression data also represent a technological breakthrough, and, to our knowledge, it is the first time high-throughput gene expression screening has been applied to corneal biology in an animal model. 

In conclusion, MRL mice showed accelerated corneal wound healing after alkali burn by means of rapid reepithelialization, reduced inflammation, and reduced fibrosis. Infiltrating neutrophils delayed corneal wound healing in B6 mice; however, B6 mice engrafted with MRL bone marrow cells retained the poor corneal wound healing of B6 mice, suggesting that the host environment rather than hematopoietic cells underlies the rapid healing phenotype. Array-based QPCR analyses showed transcriptional changes of fewer genes associated with inflammation and homeostasis or wound healing of corneal tissue, and increased expression of an anti-inflammatory gene, Socs1, and the corneal epithelial wound healing gene, Mmp1a. These results provide key insights into the cellular and molecular mechanisms involved in corneal wound healing. Future studies that focus on genetic crosses between MRL and C57BL/6 mice should help identify the genes that regulate corneal wound healing and help elucidate mechanisms that underlie the healing process. 

 Supported by National Eye Institute Grant EY13742, National Institute of Allergy and Infectious Disease Grant AI28802, National Institute of Diabetes and Digestive and Kidney Diseases Grant DK56597, and National Heart, Lung, and Blood Institute Grant HL077642.

 Submitted for publication May 3, 2005; revised July 5, 2005; accepted August 29, 2005.

 Disclosure: M. Ueno, None; B.L. Lyons, None; L.M. Burzenski, None; B. Gott, None; D.J. Shaffer, None; D.C. Roopenian, None; L.D. Shultz, 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.

 

Table 1.

 

Time Course of Corneal Reepithelialization in Mice after Alkali Burn

Strain	Day 1	Day 2	Day 3	Day 4	Day 5	Day 6	Day 7	Day 14	Day 28
B6	0/28 (0)	0/28 (0)	3/28 (11)	9/28 (32)	13/28 (46)	15/28a (54)	17/28a (61)	13/18a (72)	9/10a (90)
MRL	0/30 (0)	2/30 (7)	22/30* (73)	30/30* (100)	25/30* b (83)	26/30* c (87)	29/30* a (97)	19/20a (95)	10/10 (100)

 Data are the number of corneas completely reepithelialized/total number of corneas examined (% healing). Letters indicate number of eyes at each day that showed small recurrent corneal ulcers after initial reepithelialization: a, 1; b, 5; c, 4.

*   P < 0.01; statistically significant against B6, by the Fisher exact test.

Figure 1.

 

Time course of corneal opacity after alkali burn. The numbers of eyes (mice) observed were 28 to 30 (14–15), 28 to 30 (14–15), 28 to 30 (14–15), 18 to 20 (9–10), and 10 (5) per strain on days 1, 4, 7, 14, and 28, respectively. Data are shown as the mean ± SE. **Statistically significant by Mann-Whitney test, P < 0.01.

Figure 2.

 

Representative photographs of the anterior segments of eyes after alkali burn in B6 (A–D) and MRL (F–I) mice. Eyes of nontreated mice (A, F) and of alkali-burned mice killed on day 7 (B, G), 14 (C, H), or 28 (D, I) are shown. The B6 eyes on days 7 and 14 (B, C) showed severe keratitis and iritis that included ulceration (▾), inflammatory cell infiltration (δ), hyphema (ψ), fibrosis (#), edema (*), neovascularization (↓), and anterior lens epithelial hyperplasia (↑). Moderate inflammation is also present in the B6 eye on day 28 (D). On the contrary, the MRL eyes showed minimal inflammation and fibrosis on days 7 and 14, but they had apparent edema. Only slight edema was present in the MRL cornea on day 28 (I). Gross photographs (E, J) taken on day 28 are of the same eyes as seen in histology sections (D) and (I), respectively. c, cornea; i, iris; le, lens; ac, anterior chamber. Scale bar, 200 μm.

Figure 3.

 

Early histologic changes of the corneas after alkali burn in B6 and MRL mice. The left column (A–E) represents corneas from B6 mice, and the right column (F–J) represents corneas from MRL mice. Corneas of nontreated mice (A, F) and corneas sampled on days 1 (B, G), 2 (C, H), 3 (D, I), and 4 (E, J) after alkali burn are shown. The epithelial layer was absent in both strains on day 1 because of treatment with alkali, and regenerating epithelial cells were present in sections shown in (H), (D), (I), and (J). Keratocyte loss was present in both strains and were most plainly seen in MRL corneas (*). (↑) Keratocytes in untreated corneas. Infiltration of inflammatory cells into the corneal stroma (↓) or anterior chamber (δ) was present in B6 corneas, whereas fewer inflammatory cells were seen in MRL corneas. e, epithelial layer; s, stroma; ac, anterior chamber. Scale bar, 100 μm.

Figure 4.

 

Peripheral blood leukocyte analysis in B6 mice treated with anti-mouse Gr-1 antibody or PBS. The data are shown as the mean ± SE. There were 15 mice per group. Statistically significant by unpaired t-test at *P < 0.05 and **P < 0.01.

Table 2.

 

Effect of Depletion of Circulating Neutrophils on Corneal Reepithelialization after Alkali Burn in B6 Mice

Treatment	Day 1	Day 2	Day 3	Day 4	Day 5	Day 6	Day 7
PBS	0/29 (0)	0/29 (0)	8/29 (28)	14/29 (48)	14/29a (48)	18/29b (62)	23/29 (79)
αGr-1	0/30 (0)	0/30 (0)	12/30 (40)	23/30* (77)	23/30* c (77)	24/30 (80)	25/30 (83)

 Data are the number of corneas completely reepithelialized/total number of corneas examined (% Healing). Letters indicate the number of eyes at each day that showed small recurrent corneal ulcers after initial re-epithelialization: a, 3; b, 2; c, 1.

*   P < 0.05; statistically significant against PBS treatment by the Fisher exact test.

Figure 5.

 

Representative photographs of anterior segments of eyes after alkali burn of control (A) and neutrophil-depleted (B) B6 mice. Reduced inflammatory cell infiltration in the cornea and anterior chamber, absence of hemorrhage, and reduced fibrosis in the anterior chamber were visible in the neutrophil-depleted mice. Scale bar, 200 μm.

Figure 6.

 

Representative flow cytometry results of PBMCs for H2-K^k (x-axis) and H2-K^b (y-axis). The animals were lethally irradiated B6 mice engrafted with bone marrow cells from B6 (B6→B6) or MRL (MRL→B6) mice. The results of healthy nontreated animals (B6 Control, MRL Control) are also shown. The numbers represent percentages of single positive-staining cells.

Table 3.

 

Effect of Engraftment of Bone Marrow from Fast (MRL)-or Slow (B6)-Healing Mice on Corneal Reepithelialization after Alkali Burn

Recipient Strain	Donor Strain	Day 1	Day 2	Day 3	Day 4	Day 5	Day 6	Day 7
B6	B6	0/30 (0)	0/30 (0)	1/30 (3)	4/30 (13)	5/30 (17)	6/30a (20)	10/30 (33)
B6	MRL	0/28 (0)	0/28 (0)	0/28 (0)	2/28 (7)	4/28 (14)	7/28 (25)	9/28 (32)

 Data are the Number of Corneas Completely Reepithelialized/total number of corneas examined (% Healing). Differences at all time points are not statistically significant by the Fisher exact test.

^a  One eye showed a small recurrent corneal ulcer after initial reepithelialization.

Figure 7.

 

Representative photographs of anterior segments of eyes after alkali burn of B6 mice engrafted with B6 (A) or MRL (B) bone marrow cells. Note inflammation and fibrosis as severe as B6 chimera are present in MRL chimera. Scale bar, 200 μm.

Table 4.

 

AQPCR Analyses in Sham-and Alkali-Burned Corneas

Rank	GPR Score	Gene Symbol	Change (x-fold)	Name
A. Sham MRL vs. Sham B6 (control)
1	0.952	H28	93.53	Histocompatibility 28
2	0.444	Flt1	46.14	FMS-like tyrosine kinase 1 (VEGFR1)
B. Alkali B6 vs. Sham B6 (control)
1	0.986	Satb1	−281.08	Special AT-rich sequence-binding protein 1
2	0.943	Cd48	106.30	CD48 antigen
3	0.929	Spp1	98.70	Secreted phosphoprotein 1 (osteopontin)
4	0.844	Tyrobp	28.62	TYRO protein tyrosine kinase binding protein (Dap12)
5	0.709	Tnfsf13b	−16.67	Tumor necrosis factor (ligand) superfamily, member 13b (BAFF)
6	0.681	Icam1	12.64	Intercellular adhesion molecule
7	0.674	Inpp5d	13.11	Inositol polyphosphate-5-phosphatase D (SHIP)
8	0.652	Cd14	14.79	CD14 antigen
9	0.645	Sfpi1	12.88	SFFV proviral integration 1 (PU.1)
10	0.645	Flt1	28.62	FMS-like tyrosine kinase 1 (VEGFR1)
11	0.603	Il1r2	9.21	Interleukin 1 receptor, type II
12	0.525	Pfkp	6.97	Phosphofructokinase, platelet (PFK-C)
13	0.525	Fcgr3	23.31	Fc receptor, IgG, low affinity III (CD-16)
14	0.504	Pax6	−3.85	Paired box gene 6
15	0.482	Enpep	−37.73	Glutamyl aminopeptidase (BP-1/6C3, APA)
16	0.475	Hspa2	15.26	Heat shock protein 2 (Hsp70-2)
17	0.447	Ly6a	7.90	Lymphocyte antigen 6 complex, locus A (Sca-1)
18	0.433	Bcl2	−3.75	B-cell leukemia/lymphoma 2
19	0.418	Mmp13	18.82	Matrix metalloproteinase 13
C. Alkali MRL vs. Sham MRL (control)
1	0.863	Enpep	−67.47	Glutamyl aminopeptidase (BP-1/6C3, APA)
2	0.657	Ly6a	11.26	Lymphocyte antigen 6 complex, locus A (Sca-1)
3	0.647	Icam1	25.94	Intercellular adhesion molecule
4	0.559	H28	−10.49	Histocompatibility 28
5	0.510	Tyrobp	11.31	TYRO protein tyrosine kinase–binding protein (Dap12)
6	0.500	B2m	5.84	Beta-2 microglobulin
7	0.451	Ccr7	8.67	Chemokine (C-C motif) receptor 7 (CD197)
D. Alkali MRL vs. Alkali B6 (control)
1	0.842	Mmp1a	53.63	Matrix metalloproteinase 1a (interstitial collagenase)
2	0.783	Gp49a	−39.24	Glycoprotein 49 A
3	0.733	Tnfrsf13c	175.85	Tumor necrosis factor receptor superfamily, member 13c (BAFF-R)
4	0.567	Ccl11	6.44	Small chemokine (C-C motif) ligand 11 (eotaxin)
5	0.508	Ccl28	11.94	Chemokine (C-C motif) ligand 28 (CCK1)
6	0.500	Sfpi1	−10.25	SFFV proviral integration 1 (PU.1)
7	0.458	Socs1	11.73	Suppressor of cytokine signaling 1

 Results showing genes for which expression changed when comparing biological cohort samples after global normalization and ranked based on the GPR algorithm. Only comparisons with a GPR score of ≥0.4 are shown, indicating that the expression change between experimental cohorts was significant at P ≤ 0.05 when compared with at least 40% of the eligible normalizer genes. The numbers of eligible genes (i.e., genes with expression below 37.5 cycles in at least one cohort) and eligible normalizers (i.e., genes with expression below 37.5 cycles in both cohorts) were 167 (126), 203 (141), 162 (102), and 197 (120), in comparisons of A, B, C, and D, respectively. The number of biological replicates per cohort was three. The expression change (x-fold) was calculated with respect to the 10 best normalizers chosen by the GPR algorithm.

The authors thank the staff of the Department of Histology and Microscopy Science and Department of Microchemistry, The Jackson Laboratory, for their help; and Richard S. Smith, The Jackson Laboratory, for a critical reading of the manuscript.