Expression of Integrins and MMPs during Alkaline-Burn-Induced Corneal Angiogenesis

Abstract

purpose. To determine in a corneal alkaline burn model of angiogenesis whether the expression of integrins and MMPs is consistent with a VEGF-induced angiogenic response.

methods. Neovascularization in female Sprague-Dawley rats was induced by alkaline cauterization of the central cornea. RT-PCR for integrins α₁, α₂, β₃, and β₅; the endothelial marker CD31; and metalloproteinases MMP-2 and MT1-MMP was performed on naive corneas and on cauterized corneas 72 and 288 hours after cautery. Analyses of protein and MMP expression were conducted on naive corneas and on cauterized corneas 24, 72, 120, and 168 hours after cautery by immunofluorescence microscopy and gelatin zymography.

results. RT-PCR indicated a correlation between the induced angiogenic response and the expression of α₁ and β₃ integrin subunits and MT1-MMP. Immunohistochemical analysis indicated that α₁, α₂, α₅, and β₅ integrins and MMP-2 and MT1-MMP were expressed on the newly developing vasculature. The β₃ integrin was preferentially expressed on platelets.

conclusions. Integrin expression during neovascularization of rat corneas in response to alkaline injury correlates with an angiogenic response that uses the VEGF/α_vβ₅ pathway. MMP-2 and MT1-MMP, but not MMP-9, are expressed in a pattern consistent with their involvement in the angiogenic response.

Angiogenesis in adult tissue is the result of a complex interplay among proangiogenic factors, cell adhesion, and matrix remodeling. ¹ Inhibition or disruption of either cell adhesion or matrix-degrading enzymes, many of which belong to the family of matrix metalloproteinases (MMPs), is capable of blocking an angiogenic response. ^{2 3 4 5} The adhesion receptors and MMPs involved in an angiogenic response have been correlated with the presence of select initiating factor(s). ^{6 7 8} Induction of angiogenesis by bFGF or TNF-α is associated with the selective upregulation of α_vβ₃ and involvement of MT1-MMP and MMP-2, whereas induction of angiogenesis by VEGF, TGF-β, or PMA, is associated with selective upregulation of α_vβ₅. ^{6 7} Although these two pathways are well recognized, recent studies suggest that under pathologic conditions the correlation between growth factors and integrin expression is not always maintained. In several instances in which VEGF is present, both α_vβ₃ and α_vβ₅ are expressed, and in at least one study it was shown that the functional significance of α_vβ₃-mediated angiogenesis may reflect the presence of ligand for α_vβ₃. ^{9 10 11} However, not all aspects of angiogenesis are dependent on expression of α_vβ₃ or α_vβ₅ integrins. Knockout mice for α_v and β₃ integrins appear to undergo extensive vasculogenesis and angiogenesis, although in the α_v null, subtle vascular defects are present, resulting in both embryonic and postnatal death. ^{5 12} These results suggest that other integrin family members may compensate for the loss of α_v or β₃ integrins or may play a more essential role in the angiogenic response. Other members of the integrin family implicated in mediating an angiogenic response include α₁β₁, α₂β₁, and α₅β₁ integrins, which, like α_v integrins, have also been divided into bFGF-associated (α₅β₁) or VEGF-associated (α₁β₁, α₂β₁) angiogenic events. ^{13 14 15}  

Recently, the corneal alkaline burn model of angiogenesis has been characterized as having high levels of VEGF present during active vessel growth, suggesting that VEGF is the primary angiogenic factor within this model system. ¹⁶ Consistent with this finding, pharmaceutical intervention with α_vβ₃ antagonists has no effect on the angiogenic response, ¹⁷ suggesting that angiogenesis occurs through an α_vβ₅ adhesion pathway that is consistent with a VEGF-mediated angiogenic response. However, expression of α_vβ₅ was not established in these studies, and no other potential adhesion receptors have been identified. 

The purpose of this study was to characterize the pattern of integrin and MMP expression to determine whether it is consistent with a VEGF-mediated angiogenic response. In agreement with a VEGF-mediated angiogenic response, neovascularization was associated with expression of α_vβ₅, α₁β₁, and α₂β₁ integrins as well as α₅β₁. Preferential staining of α_vβ₅ and α₅β₁ was seen in the invasive angiogenic front, whereas α₂β₁ integrin appeared to be preferentially expressed in regions of vessel maturation. MMP-2 and MT1-MMP were associated with both vessel formation and the robust inflammatory response. These data indicate that the corneal alkaline burn model provides an in vivo model system to examine the role and involvement of α_vβ₅ integrins and MT1-MMP and MMP-2 in neovascularization of corneal tissue. 

Materials and Methods

Reagents and Antibodies

5-Bromo-2-deoxyuridine (BrdU) was purchased from Roche Molecular Biochemicals (Indianapolis, IN); RNA extraction reagent (TRIzol), reverse transcriptase (SuperScript II), and bFGF from Invitrogen (Carlsbad, CA); and gelatin zymography polyacrylamide gels (10%), renaturing buffer, and developing buffer from Novex (San Diego, CA). Primary antibodies were purchased from the following companies and used at the concentrations shown: goat anti-type IV collagen (1:250 dilution; 1.6 μg/mL; Southern Biotechnology Associates, Inc., Birmingham, AL); rabbit polyclonal anti-integrin α₁, α₂, α₃, α₅, and β₅ subunits and anti-CD31 (1:100 dilution for the α subunits 1:500 dilution for the β₅ subunit; Chemicon International, Inc., Temecula, CA); mouse monoclonal anti-rat integrin β₃ chain (1:100 dilution; 5 μg/mL; PharMingen, San Diego, CA); and rabbit polyclonal anti-MMP-2 and MT1-MMP (Chemicon). All secondary antibodies were F(ab′)2 fragments conjugated to either tetramethylrhodamine isothiocyanate (TRITC) or fluorescein isothiocyanate (FITC; both 1:200 dilution Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). 

Animal Model

Female Sprague-Dawley rats, weighing 250 to 300 g, were anesthetized with isoflurane (4% vol/vol) and topical application to the corneal surface with proparacaine 0.1% (Allergan, Inc., Irvine, CA). The alkaline burn was created by touching the central cornea with the tip of a silver nitrate applicator (75% silver nitrate, 25% potassium nitrate; Grafco; Graham-Field, Inc., Hauppauge, NY) for 2 seconds. At the indicated times, animals were killed and the eyes enucleated for various studies at postinjury intervals ranging from 24 to 288 hours. For immunofluorescence analysis, the eyes were embedded in optimal cutting temperature (OCT) solution and cryosectioned. For wholemount studies, entire corneas were removed and quartered. Experimental animals were treated and maintained in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Cryosectioning and Immunofluorescence

The eyes (injured or naive) were sagittally cryosectioned in 8- to 13-μm sections for immunostaining with mouse monoclonal or goat and rabbit polyclonal antibodies. The sections were fixed in 100% acetone for 5 minutes, briefly dried, rehydrated in phosphate-buffered saline (PBS) and incubated in a moist chamber as follows: 5% BSA (Sigma, St. Louis, MO) in PBS for 2 hours, primary antibodies for 2 hours at room temperature, five washes in PBS for 5 minutes each, secondary antibodies conjugated to fluorochromes for 1 hour at room temperature, and five more washes as before. Samples were mounted with fluorescence medium (Fluoromount G; Southern Biotechnology Associates) and observed and photographed with a compound microscope (model E800; Nikon, Tokyo, Japan) equipped with a digital camera (Spot; Diagnostic Instruments, Inc., Sterling Heights, MI). Colocalization of the angiogenesis-related molecules and vascular markers was achieved by using various combinations of mouse, goat, and rabbit primary antibodies. Negative controls for immunostaining were naive serum or purified IgG for each species of primary antibody used, as well as for secondary antibody alone. In all instances tissues were costained with collagen type IV to mark the presence of vessels and to serve as an internal positive control. All control tissues were from corneas harvested 72 hours after injury, because they provided the greatest range of cellularity. 

Wholemount Immunofluorescence

Complete fresh corneas were cut in quarters and fixed in 90% methanol and 10% dimethyl sulfoxide (DMSO) for 15 minutes at room temperature, rinsed in 1× PBS three times for 2 minutes each, blocked in 2% BSA in PBS for 4 hours, incubated in primary antibody anti-β₃ and Bandeiraea simplicifolia (BS-1) lectin (Sigma) overnight at 4°C, washed 5 times in PBS for 1 hour each, incubated in secondary antibodies conjugated to fluorochromes overnight at 4°C, and washed five times as before. Finally, corneas were flatmounted and analyzed with either the compound microscope (E800 Nikon) equipped with the digital camera (Diagnostic Instruments, Inc.) or by a confocal microscope (model TCS SP; Leica Microsystems Inc., Exton, PA). 

Gelatinase Zymography

For gelatin zymography, MMPs were extracted from corneal tissue sections. For MMP extraction and electrophoresis, corneal sections were crushed using a mortar and pestle and MMPs extracted into 1% SDS-Tris 10 mM buffer (pH 7.4). After extraction, total protein was determined using a bicinchoninic acid (BCA) assay (Pierce, Rockford, IL) and 20 μg total protein was brought to 1× SDS-PAGE buffer and samples separated on 10% SDS-polyacrylamide zymography gels without heating. Gels were developed according to the manufacturer’s instructions (Novex). Corneal sections were obtained by trimming isolated corneas, leaving a central corneal strip 2 to 2.5 cm in width. This central strip was then cut into eight slices approximately 0.8 mm in width, by stacking nine single-edge razor blades and pressing them against the flattened corneas to generate the eight corneal slices, as shown in Figure 7A . Similar sections were pooled and stored on dry ice until all slices were collected. In each experiment performed, slices were obtained from four corneas. 

In Situ Zymography

Frozen tissue sections, 4 to 8 μm in thickness, were mounted onto gelatin-coated slides (Fuji Pharmaceuticals Inc., Tokyo, Japan), incubated at 37°C in a moist chamber for 4 to 6 hours, and dried at room temperature. After drying, tissues were stained (Amido Black 10B; Sigma) for 15 minutes followed by rinsing in water and then were destained (70% methanol, 10% acetic acid) for 20 minutes. Images were captured by bright-field microscopy. 

Reverse Transcription-Polymerase Chain Reaction

Total RNA was isolated from pooled corneal tissue (four corneas per time point were pooled; pooling of samples gave repeatable responses between multiple experiments) from naive corneas and from cauterized corneas 72 and 288 hours after cautery, using the standard RNA extraction procedure outlined in the manufacturer’s protocol (TRIzol; GibcoBRL, Rockville, MD). Isolated RNA was treated with RNase-free DNase I to remove any contaminating genomic DNA. RT-PCR analysis of RNA in the absence of reverse transcriptase was used as a negative control. The total RNA was quantitated by spectrophotometry at an absorbance of 260 nm. Total RNA (1 μg) was reverse transcribed with 50 units reverse transcriptase (SuperScript II; GibcoBRL) in the presence of 2.5 μg/mL random hexamer and 500 μM dNTP for 50 minutes at 42°C, followed at 70°C for 15 minutes. One microliter of the resultant cDNA was amplified in the presence of 1 nM sense and antisense primers, 200 μm dNTP, and 3.5 units of high-fidelity enzyme mix (Expand; Roche Molecular Biochemicals). PCR conditions: initial 5 cycles, denaturing at 94°C for 15 seconds, annealing at 58°C to 55°C for 30 seconds (decrease 0.5°C each cycle), and 72°C for 30 seconds. For the remaining 27 cycles PCR conditions were 94°C for 15 seconds, 55°C for 30 seconds, and 72°C for 45 seconds. The amplified samples were then loaded at equal volumes (10 μL) onto 1.5% agarose gels. The PCR products were visualized with ethidium bromide. The primer pairs used for amplification are given in Table 1 . All PCR products were subcloned and sequenced to verify product as the target gene. Results are representative of multiple experiments. 

Corneal Micropocket Assay

A corneal micropocket assay was performed as described by Kenyon et al., ¹⁸ using 400 ng bFGF in hydron pellet beads. Briefly, female Sprague-Dawley rats, weighing 250 to 300 g, were put under general anesthesia with 200 μL xylazine (20 mg/mL), ketamine (100 mg/mL), and acepromazine, and before surgery eyes were topically anesthetized with 0.5% proparacaine. A 1-mm corneal incision penetrating halfway through the corneal stroma was made 2.5 mm from the temporal limbus, and a pocket was generated by separating the stroma from the point of incision to approximately 1 mm from the limbal vessel. A hydron bead (0.4 × 0.4 mm) containing 140 ng bFGF was implanted in the pocket. Three and 5 days after implantation of the hydron pellet, corneas were prepared for wholemount analysis. 

Results

Effect of Alkaline-Burn-Induced Corneal Angiogenesis on Message Levels of Integrins α₁ and β₃ and MT1-MMP

In the alkaline burn corneal angiogenesis model, the cornea is wounded by the application of a silver nitrate applicator to the central region of the cornea. Twenty-four hours after wounding, an acute inflammatory response is observed. Neovascularization originating from the limbal vessels and extending toward the central corneal burn is clearly visible by 72 hours. At 168 hours after cautery, neovessels extend across the entire cornea from the limbus to the central burn. ¹⁹  

RT-PCR was performed on mRNA extracted from pooled corneas to examine the expression of integrin subunits α₁, α₂, β₃, and β₅ and of MMPs MMP-2 and MT1-MMP. Corneal tissues examined included naive corneas and cauterized corneas harvested 72 hours (3 days) and 288 hours (12 days) after cautery. To correlate gene expression relative to vessel growth, levels of CD31, a marker for endothelial cells were also examined. Messages for MMP-2 and β₅ and α₂ integrins were present at all time points examined and showed no apparent increase correlated with the angiogenic response (Fig. 1) . In contrast, the expression of MT1-MMP and α₁ and β₃ integrin subunits correlated with the angiogenic response, as revealed by increased message correlating with expression of CD31 (Fig. 1) . 

Association of Integrin Subunits with the Developing Neovasculature

Analysis of integrin expression on neovessels was accomplished by using immunologic reagents that recognized individual integrin subunits. The individual subunits examined included α₁, α₂, α₅, β₃, and β₅. For analysis of expression patterns, three separate experiments were performed in which naive corneas 72 hours after harvesting and cauterized corneas 24, 72, 120, and 168 hours after cautery were analyzed. Staining was assessed in a minimum of three adjacent sections to determine pattern and localization. Nonspecific staining at each time point was assessed by the use of naive serum or purified IgG for each species of primary antibody used, as well as for the secondary antibody alone. Staining in naive corneas for each of the integrins examined is shown in Figure 2 . Most staining was for α₁, α₂, α₅, and β₅ and was limited to the corneal epithelium. Stromal staining was also observed but to a limited extent that was not readily apparent. No immunoreactivity was seen for β₃ integrin (not shown). 

Localization of integrins to neovessels is shown in Figure 3 . Integrins α₁, α₂, α₅, and β₅ were seen to stain neovessels as identified by collagen type IV staining in longitudinal sections of developing neovessel (Figs. 3A 3B 3C 3D 3E 3F 3G 3H 3I 3J 3K 3L) . Integrin β₃ appeared to stain individual cells within the lumen (Figs. 3M 3N 3O) . Positive immunoreactivity for α₁, α₂, α₅, and β₅ not colocalizing with collagen type IV was also observed. This later staining was positive on stromal fibroblast and inflammatory cells. Staining for α₁ and α₅ was also seen in the corneal endothelium. No stromal staining was observed for β₃ integrin. 

The punctate staining pattern for β₃ integrin suggests localization within the developing vasculature, potentially on platelets and not endothelium (Figs. 3M 3N 3O) . Confocal microscopy of wholemounted corneas indicated that this immunostaining is associated with expression of β₃ on platelets (Figs. 4A 4B) . To confirm that the staining pattern for β₃ is not associated with neovascularization, we examined, by corneal micropocket assay, the expression of β₃ in tissues in which corneal angiogenesis had been induced by bFGF. Expression of β₃ in the corneal micropocket assay was restricted to the leading part of the vasculature (Figs. 4C 4D) and was pronounced on endothelial cells (Figs. 4E 4F) . These results are consistent with previous studies examining α_vβ₃ expression in neovascularized tissue, ²⁰ but contrast greatly with results in the corneal alkaline-burn model, indicating that β₃ expression is principally restricted to platelets. 

Differential Distribution of α₅ and β₅ Integrins within the Developing Neovasculature

Spatial distribution of integrin staining along the developing neovessels is shown in Figure 5 . Expression of α₁ integrin within the developing vasculature showed a uniform pattern of staining throughout (Figs. 5A 5B) . Similar to staining for α₁ integrin, staining for α₅ integrin was also seen throughout the developing vasculature; however, more pronounced staining was typically seen in distal regions of the neovasculature (Figs. 5C 5D) . Staining for β₅ integrin was preferentially seen in more distal regions of the developing neovasculature associated with the invasive front (Fig. 5E 5F) . Staining for α₂ integrin was preferentially observed in regions of vessel maturation, with little or no staining in more distal regions associated with the invasive front (Figs. 5G 5H) . 

Expression of MT1-MMP and MMP-2

MT1-MMP and MMP-2 form a functional complex on the cell surface involved in the proteolytic modification of the surrounding extracellular matrix. ^{21 22} The observed increase in MT1-MMP expression during neovascularization suggests that MT1-MMP and MMP-2 may form a functional complex on endothelial cells, thus promoting the neovascular response. To determine the distribution of MT1-MMP and MMP-2, immunostaining was preformed. Both MT1-MMP and MMP-2 stained the developing vasculature (Figs. 6A) . No clear spatial distribution of either MT1-MMP or MMP-2 was observed within the neovessels (Fig. 6B) . These data indicate that both MT1-MMP and MMP-2 were expressed in association with the neovasculature. Staining of MT1-MMP and MMP-2 was also observed within the stroma but did not colocalize with collagen type IV immunostaining. Costaining of tissues with CD18, to mark inflammatory cells, and either MMP-2 or MT1-MMP indicated that most of the stromal staining was associated with inflammatory cells. 

Correlation of MMP-2 Processing with Angiogenic Responses

MMP-2 is expressed as a latent 72-kDa inactive proenzyme that is processed to a 62-kDa active form, principally through its association with MT1-MMP. ^{21 22} Thus, the presence of the 62-kDa gelatinase activity in gelatin zymography indicates activation of MMP-2. To determine whether the observed staining correlates with MMP-2 activation, processing of MMP-2 was monitored by gelatinase zymography. To correlate MMP-2 processing with vessel growth, corneas were sectioned into 0.8-mm segments (Fig. 7A) , and gelatinase activity within each segment was examined. As shown for the 72-hour time point, conversion of the 72-kDa form of MMP-2 to its 62-kDa active form was observed in segments 1 and 2 (Fig. 7B) , which corresponds with an angiogenic response. A similar correlation between neovessel formation and the presence of the 62-kDa active form of MMP-2 was also observed at the 120- and 168-hour time points. Lower levels of the 62-kDa forms of MMP-2 were also observed in segments 3 and 4, associated with wound healing within the central cornea. 

In addition to MMP-2, MMP-9 expression and activation were detected by gelatin zymography. Conversion of MMP-9 from its latent 92-kDa form to its activated 82-kDa form was seen principally in segments 2, 3, and 4, suggesting a correlation with the wound-healing response (Fig. 7B) . ^{23 24 25} Expression and conversion of MMP-9 was greatest at 24 hours after cautery and was nearly absent by 120 hours (data not shown). The pattern of MMP-9 expression did not correlate spatially with vessel development. 

Correlation of Expression and Conversion of MMP-2 with In Situ Gelatinase Activity

The pattern of MMP-2 conversion detected by gelatinase zymography represents both active enzyme and that associated with tissue inhibitors of matrix metalloproteinases (TIMPS) in an inactive complex. To identify endogenously active gelatinase within the cornea, in situ zymography was performed (Fig. 7D) . Consistent with the pattern of MMP-2 processing by gelatin zymography, the pattern of gelatinase activity observed reflected a gradient of gelatinase activity with highest levels in the limbal region and lower levels in adjacent tissue within the corneal stroma (Fig. 7D) . The pattern of gelatinase activity in the cornea was consistent with regions of neovascularization. Strong punctate gelatinase activity was also seen in the corneal stroma, which probably reflects gelatinase activity on inflammatory cells. 

Discussion

We examined the pattern of integrin and MMP expression within the corneal alkaline burn model relative to the angiogenic response by RT-PR, immunohistology, and gelatinase zymography. A summary of the immunohistochemical and zymography analysis for the integrins and MMPs studied are presented in Table 2 . Analysis of integrin and MMP expression by RT-PCR demonstrated that expression of CD31, integrins α₁ and β₃, and MT1-MMP correlated with the angiogenic response. No alteration was detected in the levels of α₂ and β₅ integrins and MMP-2 that was related to neovascularization. The inability to detect a change in message for α₂ and β₅ integrins and for MMP-2 probably reflects the existence of abundant message present in naive tissues within the corneal epithelium for β₅ and α₂ integrins ²⁶ or within the corneal stroma for MMP-2. ²⁷  

Immunohistochemical analysis of α₁, α₂, α₅, and β₅ subunits indicated expression of these subunits within the developing neovessels. The α₁ integrin was uniformly expressed within the developing neovasculature, whereas α₂ appeared to be more prevalent in regions of vessel maturation. The α₅ and β₅ integrin subunits showed a preferential localization to the more distal regions of vessel formation associated with the invasive and early remodeling phases of vessel development. 

Expression of the β₃ integrin subunit was principally restricted to platelets within the developing vasculature. Staining on platelets and not on endothelial cells was confirmed by comparing β₃ staining from the corneal burn model with β₃ staining induced by bFGF in the corneal micropocket assay. The presence of a β₃-specific band in the RT-PCR analysis may represent the expression of α_vβ₃ on macrophages, ²⁸ which are present as part of the inflammatory response in this model system. Alternatively, the β₃ mRNA message detected by RT-PCR may be the result of expression on endothelial cells, which showed a low level of staining localized to the vessel lumen. 

Expression of integrin was analyzed in this study with immunologic reagents directed against individual integrin subunits. In most cases, the staining reflected the presence of a heterodimer pair, because α₁, α₂, and α₅ pair only with the β₁ integrin subunit and β₅ pairs only with the α_v subunit—the only exception being the β3 integrin subunit, which heterodimerizes with α_v and α_iib forming α_vβ₃ and α_iibβ₃ heterodimer pairs. ²⁹ We were able to exclude the presence of α_vβ₃, because the primary staining was associated with platelets that express only the α_iibβ₃ heterodimer. 

The other aspect of angiogenesis studied was the expression of MMP-2 and -9 and MT1-MMP. MMP-2 and MT1-MMP were observed to be associated with the angiogenic response based on both zymographic and immunohistochemical analyses. The correlation between the 62-kDa form of MMP-2 and MT1-MMP immunoreactivity suggests that MT1-MMP is associated with the conversion of the 68-kDa form of MMP-2 to its 62-kDa form in this model system; however, other mechanisms of MMP-2 processing may also be present. ^{30 31} Currently, MMP-2 and MT1-MMP are believed to form a functional complex in conjunction with α_vβ₃ and TIMP-2 on the cell surface, which in turn mediates localized pericellular proteolysis of the extracellular matrix, which is essential for endothelial cell migration and invasion. ^{32 33 34} In the alkaline-burn-induced corneal angiogenesis model, α_vβ₃ does not appear to play a major role in mediating the angiogenic response, and thus the role of MT1-MMP and MMP-2 within this model may be outside their association with α_vβ₃. Recently, MT1-MMP has been shown to directly mediate cell migration and adhesion and processing of integrin α chains. ^{35 36} This may be an alternate pathway through which MT1-MMP participates in regulation of cellular responses, outside the formation of a complex with α_vβ₃. In addition to MMP-2 and MT1-MMP, we observed increased levels of both the 92- and 82-kDa forms of MMP-9. The temporal and spatial patterns of MMP-9 expression both suggest its association with wound healing and not with neovascularization. ^{25 37 38 39}  

In conclusion, the α_vβ₅ integrin appears to be the principal α_v integrin associated with endothelial cells within the corneal alkaline burn model of inflammation-mediated angiogenesis. In addition to β₅, the α₁, α₂, and α₅ integrins showed consistent localization to the developing vasculature bed. Of particular interest was the preferential localization of α₅ to more distal regions of the developing vasculature and the preferential expression of α₂ integrin to regions of vessel maturation. Both MT1-MMP and MMP-2 were seen in association with the neovasculature, and the conversion of MMP-2 from its 72-kDa latent form to its 62-kDa active form correlated with neovascularization. Based on the characterization of integrin subunit expression, the data suggest that therapeutic approaches to inhibition of corneal angiogenesis in response to an alkaline burn would be best if directed against the β₅ integrin subunit and not against β₃. 

 Submitted for publication April 18, 2001; revised November 9, 2001; accepted November 30, 2001.

 Commercial relationships policy: E.

 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 7.

 

Gelatinase zymography of neovascularized corneal tissue. (A) Schematic of corneal segments analyzed by gelatinase zymography in (B). (A, dashed lines and numbers): segments used for MMP analysis; solid lines: extent of vessel growth 72 hours after cautery. (B) Analysis of MMP activity in corneal segments depicted in (A) from naive and corneas 72 hours after cautery. Numbers correlate with sections labeled in (A) with 1 correlating with the limbal regions and 4 with the central corneal wound (A, W). (C) Gelatinase activity was inhibited by EDTA but not by a general protease inhibitor cocktail. Con: zymography developed under standard conditions used for analysis of MMP activity; EDTA: zymography developed in the presence of EDTA showing inhibition of MMP activity; Inhib: zymography developed in the presence of protease inhibitor cocktail. Lane 1, sample from 24-hour tissue segment 4 in which high levels of MMP-2 and -9 were seen; lane 2, MMP-2 and -9 standards. (D) In situ zymography of frozen corneal sections from (A) naive corneas and from cauterized corneas (B) 72 and (C) 120 hours after cautery. MMP activity, indicated by the absence of staining, was associated with neovessels, and was present in the limbus (B, C, arrows and arrowheads, respectively). Pronounced MMP activity was also observed in individual cells throughout the cornea 72 hours after cautery (B, ★).

Table 1.

 

Oligonucleotide Primer Sequences

Primer	Oligonucleotide Sequence	Fragment Size (bp)
MT1-MMP	5′-GTGACAGGCAAGGCCGATTCG-3′	446
	5′-TTGGACAGTCCAGGGCTCAGC-3′
MMP-2	5′-ACTCCTGGCACATGCCTTTGCC-3′	401
	5′-TAATCCTCGGTGGTGCCACACC-3′
Integrin β3	5′-TTTGCTAGTGTTTACCACGGATGCCAACAC-3′	866
	5′-CCTTTGTAGCGGACGCAGGAGAAGTCAT-3′
Integrin β5	5′-CGAATGGCTGTGAAGGTGAGATTGA-3′	854
	5′-CAGTGGTTCCAGGTATCAGGGCTGTAAAAT-3′
Integrin α2	5′-CAAGCCTTCAGTGAGAGCCAAGAAACAAAC-3′	728
	5′-CAAACC TGCAGTCAATAGCCAACAGGAAAA-3′
Integrin α1	5′-GGAGAACAGAATTGGTTCCTACTTTGG-3′	335
	5′-CGGAGCTCCWATCACGAYGTCATTAAATCC-3′
CD31	5′-GGCATCGGCAAAGTGGTCAAG-3′	680
	5′-CAAGGCGGCAATGACCACTCC-3′
Actin	5′-ATCTGGCACCACACCTTCTACAATGAGCTGCG-3′	837
	5′-CGTCATACTCCTGCTTGCTGATCCACATCTGC-3′

 W = A or T; Y = C or T.

Figure 1.

 

RT-PCR analysis for CD31, MT1-MMP, MMP-2, and integrin chains αβ₁, α₂, β₃, and β₅ from total RNA pooled from four individual corneas representing naive corneas (lane 1), and cauterized corneas 72 hours (lane 2) and 228 hours (lane 3) hours after cauterization. In naive and neovascularized corneas, message for α₂ and β₅ integrins and MMP-2 were detected. Message for α₁ and β₃ integrins, CD31, and MT1-MMP were detected only in neovascularized corneas.

Figure 2.

 

Staining in naive cornea for integrins in the limbal region (A, C, E, G) and central cornea (B, D, F, H): (A, B) α₁, (C, D) α₂, (E, F) α₅, and (G, H) β₅.

Figure 3.

 

Localization of α₁, α₂, α₅, β₅, and β₃ to neovessels 120 hours after initiation of a neovascular response. (A, D, G, J, M) Staining for the individual integrins; (B, E, H, K, N) staining with collagen type IV to mark neovessels; (C, F, I, L, O) merged images for both integrin and collagen type IV staining. (A, B, C) Staining for α₁, (D, E, F) for α₅, (G, H, I) for β₅, (J, K, L) for α₂, and (M, N, O) for β₃.

Figure 4.

 

Wholemount analysis of β₃ integrin staining in cautery model (A, B) or in bFGF micropocket assay (C–F). (A, B) Wholemount confocal microscopy for BS-1 lectin (TRITC) and β₃ integrin (FITC), showing localization of β₃ signal to platelets within the neovasculature (A, arrow) or at the ends of neovascular buds (B, arrow). (C–F) Wholemount immunofluorescence of BS-1 lectin (TRITC) and β₃ integrin (FITC) in neovascularized corneas induced by bFGF. (C) BS-1 lectin and (D) β₃ integrin in neovascularized cornea showing preferential distribution of β₃ integrin staining with the angiogenic front adjacent to the hydron pellet (P). (L) Limbal vessels. Confocal images of β₃ immunostaining within the neovasculature (E) and costaining of β₃ (FITC) with BS-1 lectin (TRITC). Original magnification, ×64.

Figure 5.

 

Composite images of corneas 120 hours after cautery show spatial distribution for integrins α₁ (A, B), α₅ (C, D), β₅ (E, F), and α₂ (G, H). (A, C, E, G) Staining for the individual integrins; (B, D, F, H) costaining with collagen type. Corneas are oriented with the limbal region on the left with vessel growth going from left to right toward the central corneal wound (not shown). (G, H, ★) Low level of α₂ staining in distal regions of the neovasculature. (G, arrow) Pronounced α₂ associated with inflammatory cells.

Figure 6.

 

Staining for MMP-2 and MT1-MMP in neovascularized cornea. (A) Localization of MMP-2 and MT1-MMP to neovessels. MMP-2 staining (A–C), MT1-MMP staining (D–F). Arrows: localization of MMP-2 or MT1-MMP to neovessels; (★) staining not associated with neovessels. (B) Composite images of MMP-2 and MT1-MMP showing spatial distribution. (A) Staining for MMP-2; (B) coimmunofluorescence of MMP-2 with collagen type IV; (C) staining for MT1-MMP; and (D) coimmunofluorescence of MT1-MMP with collagen type IV. Images are from corneas harvested 72 hours after cautery.

Table 2.

 

Summary of Expression Patterns of Integrin Subunits and MMPs Related to Neovascular Response

Integrin Subunit/MMP	RT-PCR Naive	RT-PCR Neovascular	Neovascular Localization
α1	−	+	Throughout neovasculature
α2	++	++	Within neovasculature, with lower levels in the invasive front
α5	++	++	Throughout neovasculature, with higher levels in the invasive front
β5	++	++	Preferentially in the invasive front
β3	−	+	On platelets within neovasculature
MMP-2	++	++	Throughout cornea, conversion of 72-kDa to 62-kDa active form associated with neovascularization and wound healing
MT1-MMP	−	++	Throughout neovasculature
MMP-9	NA	NA	Expression of 92-kDa form and activation to 82-kDa form associated with wound healing