FasL-Fas Interactions Regulate Neovascularization in the Cornea

Abstract

purpose. Neovascularization of the avascular cornea is a significant problem associated with many corneal diseases. Because Fas ligand (FasL) is highly expressed in the cornea, the role of this molecule in controlling corneal neovascularization was examined in this study.

methods. C57BL/6(B6), FasL (CD95L)-deficient B6-gld, and Fas (CD95)-deficient B6-lpr mice were subjected to the suture model of neovascularization. Corneas were evaluated for neovascularization and representative samples subjected to immunohistochemical analysis for expression of Fas antigen and CD31 (platelet-endothelial cell adhesion molecule [PECAM-1]) on vessels that were present in the tissue. Corneas were also explanted and placed in collagen gel cultures to test the ability of anti-Fas antibody to prevent vessel extension from explanted corneas.

results. Immunohistochemical data demonstrated that quiescent vessels express CD31 alone, whereas vessels that penetrate the cornea coexpressed both the Fas antigen and CD31. A significant increase was observed in neovascularization in FasL-deficient B6-gld corneas compared with B6 corneas, and new vessel growth in both B6 and B6-gld was inhibited by anti-Fas antibody. Whereas Fas-deficient B6-lpr corneas displayed significantly less neovascularization than normal B6, B6-lpr mice express Fas on growing vessels. In corneal explant cultures, vessel growth from B6 and lpr mice corneas was inhibited by anti-Fas antibody, confirming functional Fas expression in B6-lpr mice.

conclusions. These data indicate that FasL is an important factor in controlling corneal neovascularization.

The formation of new blood vessels (angiogenesis) is essential for growth and development of multicellular organisms. It requires the sprouting and migration of endothelial cells, as well as endothelial cell proliferation and capillary tube formation (differentiation). This process requires extracellular proteolytic activity to degrade extracellular matrix and basement membrane for endothelial cell invasion and capillary morphogenesis to take place. ^{1 2} The growth of new vessels from preexisting blood vessels (neovascularization) is also a significant component of organ and tissue homeostasis. An inflammatory response, ³ ischemia, ⁴ or local production of angiogenic factors ^{4 5 6 7 8 9 10} can initiate this type of new vessel growth. Cell adhesion through matrix-binding integrins and proteolytic activity are also important in this process. ¹  

Although neovascularization is an essential biological process, there are many instances in which the growth of new vessels can have pathogenic consequences. This is the case in the eye, where neovascularization is a major component in several ocular disorders including diabetic retinopathy (DR), retinopathy of prematurity (ROP), and age-related macular degeneration (AMD). ⁴ In these disorders growth of new vessels can impair vision and threaten the quality of life. Many chronic corneal diseases also involve neovascularization. Growth of blood vessels into this normally avascular area compromises structural and functional integrity. When corneal allografts are placed on a vascularized graft bed, their acceptance is markedly reduced. ^{11 12 13} In fact, treatment with agents that reduce neovascularization can lead to increased acceptance of the corneal graft. ^{13 14}  

Recent studies from our laboratory have shown that the presence of FasL in the eye is a barrier to both inflammatory cells ¹⁵ and new blood vessels. ¹⁶ The control of inflammation is known to be a component of the immune privilege of the eye. ^{15 17} FasL expressed on ocular tissues induces apoptosis in Fas⁺ lymphoid cells that invade the eye in response to viral infection ¹⁵ or corneal grafting. ¹⁸ We have also found that FasL expressed in the retina controls new vessel growth beneath the retina, ¹⁶ by inducing apoptosis in vascular endothelial cells, which are known to express the Fas antigen. ^{19 20 21} The loss of FasL expression in this region may be a predisposing factor in AMD, allowing vessels to localize beneath the retina after penetration of Bruch’s membrane. This can lead to retinal detachment and visual loss. Because FasL is highly expressed in the cornea, ^{15 18 22} we performed studies to determine whether FasL might play a role in corneal neovascularization. Our results demonstrate that Fas-FasL interaction plays a significant role in regulating extension of new blood vessels into the cornea. 

Materials and Methods

Mice

Female C57BL/6(H-2^b) mice were purchased from National Cancer Institute (Frederick, MD). Female and male C57BL/6-gld and C57BL/6-lpr ²³ were initially purchased from The Jackson Laboratory (Bar Harbor, ME) and bred in our mouse facility. All mice were used at 7 to 10 weeks of age. These mice are housed in pathogen-free conditions in the Washington University Department of Comparative Medicine (DCM) facilities in adherence to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Antibodies

Hamster anti-mouse Fas (JO-2) was purchased from BD Pharmingen, Inc. (San Diego, CA) for use in the in vitro blocking assays. For immunohistochemical staining studies we purchased rabbit anti-Fas (M20) conjugated with FITC (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). We purchased rat anti-CD31 (platelet-endothelial cell adhesion molecule [PECAM-1], clone 390) from BD PharMingen Inc. Secondary antibody mouse anti-rat IgG Cy3-labeled was purchased from Jackson ImmunoResearch (West Grove, PA). 

Corneal Sutures

Three interrupted sutures were put into the corneal as previously described. ²⁴ Briefly, both donor and recipient mice were anesthetized by intraperitoneal injection of ketamine and xylazine (100 mg/mL of 1.0 mL ketamine, 100 mg/mL of 0.15 mL xylazine, 4.60 mL dH₂O with 0.1 mL/25 g mouse body weight). Tropicamide (1% Mydriacyl; Alcon, Humacao, Puerto Rico) was used to dilate the pupil before suturing. The central cornea of each eye received three sutures, using 11-0 monofilament nylon (70-μm diameter needle; Alcon, Ft. Worth, TX). Antibiotic ointment (Ocumycin; Pharmafair, Hauppauge, NY) was placed on the corneal surface for 2 days after the surgery to limit infection. Sutured eyes from each mouse were examined by two masked observers and scored for opacity, on a scale of 0 to 5, and neovascularization, on a scale from 0 to 8, from day 9 to day 24 after suturing, as previously described, ^{24 25} using a vertically mounted slit lamp biomicroscope (Marco, Jacksonville, FL). 

Pellet Preparation and Implantation

Sucralfate pellets were prepared and inserted into micropockets in the corneas of mice as previously described. ²⁶ Briefly, ethanol-based Hydron solution (12%, Interferon Sciences, New Brunswick, NJ) was mixed with either anti-Fas (JO-2) or bovine serum albumin (BSA) along with 10 mg sucralfate (Marion Merrell Dow, Cambridge, MA) and air dried on a nylon mesh (Small Parts, Inc., Miami Lakes, FL) to form pellets. This procedure resulted in pellets of less than 5 μl, which were implanted into the cornea of anesthetized mice. Implantation was performed using a dissecting microscope by making an incision across the eye with a scalpel blade, then forming an intrastromal pocket with a modified iris spatula (FST, Foster City, CA). A pellet was then inserted into the pocket using fine surgical forceps (FST). The pocket flap was re-apposed and ophthalmic antibiotic ointment (E. Fougera & Co., Melville, NY) applied to prevent infection and provide lubrication to the cornea. 

Immunohistochemical Staining of Sutured Corneas

Sutured mouse corneas were surgically removed and quick frozen in OCT compound (Miles Inc., Elkhart, IN) with liquid nitrogen. The corneas were processed into 40-μm sections, mounted on microscope slides, and air dried for 30 minutes. After fixation in methanol at −20°C for 7 minutes, they were allowed to air dry for 1 hour. Sections were rinsed twice in PBS and blocked with 5% normal goat serum and 0.3% Triton-X in PBS for 30 minutes at room temperature. Sections were then sequentially stained with anti-CD31 for 2 hours at room temperature followed by overnight incubation with mouse anti-rat IgG-Cy3 at 0°C. The slides were then stained with polyclonal anti-Fas conjugated with FITC for 2 hours at room temperature. After staining the slides was analyzed using a confocal microscope (LSM-410; Carl Zeiss, Thornwood, NY). The 8-bit confocal images were analyzed with image analysis software package (Metamorph; Universal Imaging Corp., Downingtown, PA). 

Neovascularization Assay in Collagen Gel

Four sutures (10-0 nylon) were placed into the corneas. Ten to 14 days later, mice that displayed significant corneal vessel growth (scores >5) were perfused with 10 mL Ca⁺²-, Mg⁺²-free HBSS through the heart. The central cornea was removed and cut into quarters, and each section was placed into a single well of a 24-well plate on a thin collagen gel matrix. ²⁷ The collagen gel was prepared by mixing 10× PBS with 1 N NaOH and 2.4 mg/mL rat-tail collagen, type I (BD Biosciences, Bedford, MA). Each well received 250 μL of the collagen gel, which was incubated at 37°C for 30 minutes to allow the gel to solidify. After placement of the corneal specimen, it was overlaid with 1.5 mL DMEM containing 10% fetal bovine serum. In some experiments medium contained 0.1 mg/mL anti-Fas antibody (JO-2). Plates were incubated at 37°C for 10 to 15 days. During the incubation period, the development of vascular tubes extending from the cornea into the surrounding collagen gel was monitored by microscope. 

Statistical Analysis

Neovascularization and opacity scores for each group in an experiment were compared by a Kaplan-Meier log rank test to determine the statistical significance for differences between groups. Results were then confirmed by a Wilcoxon analysis. 

Results

A preliminary study revealed that corneal neovascularization was maximal in C57BL/6(B6) mice by 10 to 14 days (data not shown) after suturing. Therefore, on day 14 we examined corneas for coexpression of the endothelial cell marker CD31 and Fas. Results showed that the CD31⁺ vessels within the cornea coexpressed the Fas antigen (Fig. 1) . In contrast, preexisting vessels found in the retina expressed only CD31 and not Fas antigen (Fig. 2) . Because it is well known that the cornea expresses FasL, ¹⁸ our observations suggest that expression of FasL by the cornea may be positioned to regulate the extent of growth of new blood vessels in the cornea. 

We next examined the growth of new vessels after the placement of sutures in the corneas of B6, B6-gld (FasL⁻), and B6-lpr (Fas⁻) mice. Results show that corneas of B6-gld animals had significantly higher scores than did B6 corneas (Fig. 3) . Neovascularization in B6-lpr mice was significantly less than in normal B6 mice (Fig. 3) . Because several recent reports have suggested that the lpr mutation is somewhat “leaky,” ^{28 29} we stained the growing vessels in these vascularized corneas for Fas. Results in Figure 4 show that developing blood vessels in the B6-lpr cornea stain with anti-Fas and anti-CD31. Lymphoid cells in these lpr mice were negative for Fas staining (data not shown). 

Results thus far are consistent with the idea that FasL expression may regulate the growth of new vessels in the cornea; however, we decided to test our hypothesis more directly in vitro. To do this, we developed an in vitro assay in which vascularized corneas were cultured on a collagen gel matrix. In these cultures, vessels extend from the cornea into the surrounding collagen matrix (Fig. 5) . These growing vessels expressed both Fas and CD31 (data not shown). When anti-Fas antibody (JO-2) was added to the gel at the initiation of the culture the formation of vessels by the cornea was prevented (Fig. 5) . When corneas from B6-lpr mice were used in this assay, anti-Fas also prevented vessel growth (Fig. 5) . This confirms the observations shown in Figure 4 and demonstrates that Fas expressed on developing vessels in the B6-lpr mouse was functional. 

Because anti-Fas antibody blocked vessel growth in the collagen gel assay, we tested whether similar results could be obtained in vivo. To do this, we implanted pellets (Hydron; Interferon Sciences) containing anti-Fas antibody into the corneas of B6 and B6-gld mice 5 days after suturing. Neovascularization scores over the next 4 weeks indicated that B6-gld corneas that received anti-Fas-containing pellets showed significantly less neovascularization during peak corneal vessel growth (days 16–19) than did the control (Fig. 6) . This is illustrated in photomicrographs of sutured eyes of B6-gld mice in which BSA-containing pellets were implanted (Fig. 6B) and those in which pellets containing the anti-Fas antibody JO-2 were implanted (Fig. 6C) . Note the presence of extensive growth of vessels in Figure 6B compared with that seen in Figure 6C , indicating that anti-Fas significantly inhibited vessel growth in the sutured cornea. A similar pattern of neovascularization was noted in normal B6 mice treated in the same way (data not shown). 

Discussion

The functional integrity of the cornea depends on its ability to maintain clarity. Maintaining avascularity by the exclusion of blood vessels is an important aspect of this process. When blood vessels penetrate the cornea in response to infection or physical trauma, vision can be impaired. In this article we describe the role of Fas and FasL in controlling blood vessel growth in the cornea. In our study, the interaction of FasL with Fas on the extending vessel played an important role in regulating the extent of growth of blood vessels into the central cornea. 

The cornea is considered an immune-privileged site, in part because of the expression of FasL on the endothelial and epithelial layers. ^{15 18 22} Indeed, the success of corneal transplantation in mice and humans relies on the strategic expression of FasL that prohibits lymphoid cells from damaging the tissue. ¹⁸ Recently, we have extended the concept of the protective effect of FasL against neovascularization in the retina. ¹⁶ In those studies, we demonstrated that the FasL⁺ retinal pigment epithelial (RPE) cells control the growth of new Fas⁺ blood vessels from the vascularized choroid into the eye. Because growth of new vessels in this area is a significant cause of visual loss in AMD, ⁵ these studies had important implications for understanding the pathogenesis of this ocular disorder. 

Our current results show that whereas quiescent vessels did not express Fas, vessels extending into the cornea were positive for this molecule. Corneas that did not express functional FasL (gld) showed significantly greater neovascularization than normal corneas. In addition, engagement of Fas on vessels growing in vitro prevents vascular extension. Taken together, our results suggest that FasL regulates neovascularization by engaging Fas on growing vessels and inducing apoptosis of the Fas⁺ vascular endothelial cells. This theory is supported by the recent publication by Wigginton et al. ³⁰ who demonstrated that a functional Fas/FasL interaction is required for the antiangiogenic effects of IL-12 and -2 when treating murine renal carcinoma. 

Although recent studies suggest that Fas-FasL interactions may be important in limiting the extent of growth of new vessels, ¹⁶ it is unlikely that it is involved in angiogenesis that accompanies normal development, because the vasculature of the eye (and other organs) appears to develop normally in gld and lpr animals. In addition, gld and lpr mice do not have a high incidence of spontaneous neovascularization in the eye, suggesting that other factors may be involved in controlling angiogenesis in these animals. This is analogous to what we observed with the immune privilege of the eye, where FasL works in concert with other inhibitory agents to control the spread of inflammation. The importance of FasL was only apparent when the eye was challenged with an agent that induced inflammation. ^{15 17} Similarly, we observed increased neovascularization only in the cornea (and retina ¹⁶ ) when an agent that induced angiogenesis was present. Therefore, we suggest that the Fas-FasL interaction is probably one component of a complex process. Recently, an inhibitor responsible for the avascularity of ocular compartments was identified in the cornea as pigment epithelium-derived factor (PEDF). ³¹ This protein has been shown to have neurotrophic activity ^{32 33} but is now known as a potent antiangiogenic molecule. It seems to be a constitutive component of ocular compartments, and neutralization of its activity permits new vessel growth into the central cornea. ³¹ Because apoptosis in endothelial cells is associated with the activity of PEDF, ^{34 35} we speculate that PEDF works through the action of Fas and FasL in the cornea to inhibit spontaneous neovascularization and limit induction of angiogenesis. 

Although our data suggest that FasL regulates growth of vessels into the cornea, the data also show that FasL is not an absolute barrier. Normal mice with intact Fas and FasL show development of new vessels with the appropriate stimulus. Thus, regulation of vessel growth by FasL is probably a balance between positive and negative factors. Our observations in the lpr mouse further emphasize this point, in that Fas-defective lpr mice showed significantly reduced neovascularization compared with normal mice. Because growing vessels in both normal and lpr mice express functional Fas, we hypothesize that the reduced Fas in the lpr coupled with normal levels of FasL on the cornea lead to increased inhibition (apoptosis) of growth of blood vessels. The “leakiness” of the lpr mutation ^{28 29} provides functional Fas and causes an imbalance in which the proportionally higher amounts of FasL in the lpr cornea successfully control formation of new vessels. This effect should be studied further. 

Another possible explanation of our results in lpr mice is that engagement of Fas, under some conditions, also promotes formation of vessels. This theory is supported by a recent report showing that implantation of a synthetic basement membrane (Matrigel; BD Biosciences, Mountain View, CA) containing anti-Fas antibody stimulates vessel growth. ³⁶ The report suggested that engagement of Fas in the skin stimulates angiogenesis. Thus, when Fas is defective and is expressed in significantly reduced amounts, little or no stimulation occurs. These ideas are currently under study. 

 Supported in part by National Eye Institute Grants R01 EY12707 (PMS), R01 EY06765 (TAF), and EY12826 (TAF) and Department of Ophthalmology Core Grant EY02687; an Arthritis Foundation Biomedical Science Grant, Atlanta, Georgia (PMS); a grant-in-aid from Prevent Blindness America, New York, New York (PMS); and a Foundation for Fighting Blindness research grant, Owings Mill, Maryland (TAF).

 Submitted for publication March 26, 2002; revised July 18, 2002; accepted July 19, 2002.

 Commercial relationships policy: N.

 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.

 

Expression of Fas (CD95) on new vessels in the cornea. Sutured corneas were subjected to immunofluorescent staining with antibodies to CD31 (red) and Fas (green). Overlay: merger of the two images. Magnification, ×200.

Figure 2.

 

Fas antigen was not expressed on preexisting vessels in the eye. Eyes from mice with sutured corneas were sectioned and subjected to immunofluorescent double staining with antibodies to CD31 (red) and Fas (green). Overlay: the retinal portion of the eye; merged images of staining with antibodies to CD31 and Fas. Magnification, ×200.

Figure 3.

 

Mice that did not express functional FasL demonstrated significantly more neovascularization than did normal mice. Corneas of B6 (n = 12), B6-gld (n = 10), and B6-lpr (n = 15) mice were sutured and their eyes scored for neovascularization from 9 to 24 days. Data points represent the mean neovascularization scores ± SEM in each group. Scores in B6-gld mice were higher than in other groups at all time points monitored (P < 0.01), and scores in B6-lpr mice were significantly lower than in B6 control animals on days 11, 18, and 24 (P < 0.01).

Figure 4.

 

Fas antigen was expressed by growing vessels in the Fas-deficient C57BL/6-lpr mouse. Sutured corneas were subjected to immunofluorescent staining with antibodies to CD31 (red) and Fas (green). Overlay: merged images of staining with antibodies to CD31 and Fas. Magnification, ×200.

Figure 5.

 

Anti-Fas antibodies prevented extension of vessels from vascularized corneas into the collagen gel matrix. Corneas from B6 and B6-lpr mice were placed in a collagen gel matrix for 12 days in the presence or absence of 0.1 μg/mL anti-Fas antibody (JO-2). Magnification, ×200.

Figure 6.

 

Anti-Fas antibodies reduced corneal neovascularization in vivo. (A) Neovascularization scores in sutured corneas from B6-gld mice that had pellets implanted with the JO-2 anti-Fas antibody (n = 11) and B6-gld mice that had pellets implanted that contained BSA (n = 6). Data are the mean neovascularization scores ± SEM for each group of mice. Statistical analysis revealed significant differences at days 16 (P < 0.01) and 19 (P < 0.001). It should be noted that implantation of JO-2-containing pellets in the absence of sutures did not cause neovascularization of any cornea during the 24-day observation period (n = 8). Sutured cornea from B6-gld mouse in which pellets containing (B) BSA or (C) JO-2 were implanted. (C, Arrow) The only vessels growing in this cornea. Photomicrographs were taken 17 days after placement of sutures in the animal’s cornea. Magnification, ×25.

The authors thank John Herndon and Cheryl Shomo for excellent technical assistance and Belinda McMahan of the histology core facility for preparing the tissue sections.