Nitric Oxide Synthase–II Is Expressed in Severe Corneal Alkali Burns and Inhibits Neovascularization

Abstract

purpose. Inducible nitric oxide synthase (NOS-II) is expressed in many inflammatory conditions. The implication of nitric oxide (NO) in angiogenesis remains controversial. The role of NOS-II and its influence on angiogenesis in corneal neovascularization is unknown and was investigated in this study.

methods. A mouse model of corneal neovascularization induced by chemical cauterization was used. NOS-II mRNA expression was analyzed by reverse transcriptase–polymerase chain reaction, and NOS-II protein was studied in situ by immunohistochemical analysis of the cornea. The influence of NOS-II on neovascularization was determined by comparison of vessel development in “normal” wild-type mice and mice with a targeted disruption of the NOS-II gene.

results. NOS-II mRNA was induced to very high levels after corneal cauterization and remained upregulated throughout the disease. Migratory cells in the center of the cauterization area expressed NOS-II protein. The neovascular response in mice lacking the NOS-II gene was significantly stronger than in wild-type mice, and the difference increased over time.

conclusions. These data are the first evidence that NOS-II is expressed in this model of sterile corneal inflammation. NOS-II expression inhibited angiogenesis in severe corneal alkali burns.

Nitric oxide (NO) is a free radical, synthesized from L-arginine by NO synthase (NOS). ¹ Nitric oxide is involved in diverse processes such as neurotransmission, vasodilatation, host defense, and inflammation. ^{2 3 4} The existence of at least three different forms of NOS, each coded by a specific gene, has been demonstrated. Two isoforms (the constitutive NOSs) are continuously expressed: NOS-I (nNOS), which is essentially present in neurons of the central and peripheral nervous system, and NOS-III (eNOS), which was first found in the plasmic membrane of vascular endothelial cells. ^{1 2} Both enzymes are calcium and calmodulin dependent. However, the inducible isoform NOS-II (iNOS) is calcium and calmodulin independent and is expressed in different cell types only after transcriptional activation by endotoxins or cytokines. ⁴ NO produced by NOS-II acts as an antimicrobicidal and antiviral agent ^{4 5} in immunologic defenses. Nitric oxide is also thought to be a mediator of autoimmune and inflammatory responses and is known to influence angiogenesis in many models (see below). 

Angiogenesis is an important process in physiological and pathophysiological situations. ⁶ Neovascularization is a common feature of inflammatory, infectious, and traumatic diseases of the cornea. ⁷ It is a severely disabling condition, which decreases visual acuity and worsens the prognosis of a potentially curing keratoplasty. 

The role of NO in angiogenesis can be proangiogenic, ^{8 9 10 11 12} or antiangiogenic, ^{13 14 15 16} depending on the model used. This variability may be due to the presence of different NOS isoforms in the studied models, notably NOS-II and NOS-III. It is often difficult to deduce the role of a specific form of NOS in these models because nonspecific inhibitors have been used. As far as the potential involvement of NO in corneal neovascularization is concerned, experiments with the Corneal Pocket Assay have demonstrated that NO participates in the angiogenic response to prostaglandin E₁ (PGE₁), substance-P (SP), or vascular endothelial growth factor (VEGF). ^{9 17} In our study we used the previously described model of quantified corneal neovascularization by chemical cautery, ^{18 19} chosen because of its reliability and pathophysiological relevance, because corneal alkali burns still cause considerable problems in humans. We investigated the influence of NOS-II on corneal inflammation and on angiogenesis by comparing the reaction in mice lacking the NOS-II gene and wild-type mice. 

Materials and Methods

Animals

C57BL/6x129SvEv mice with a targeted disruption of the NOS-II gene (knockout NOS-II), generated as previously described, ²⁰ were generously provided by John D. MacMicking, Carl F. Nathan (Cornell University Medical College, New York), and John S. Mudgett (Merck Research Laboratories, Rahway, NJ). NOS-II–deficient mice were mated with C57BL/6x129SvEv wild-type (+/+) mice to produce heterozygous (+/-) NOS-II–deficient mice. These heterozygotes were then mated to provide NOS-II–deficient mice (-/-) and wild-type littermates (+/+) with the same genetic background. The animals used for these experiments were between generations 8 and 15, bred continuously from these NOS-II knockout and wild-type littermates. Genotyping to verify the lack or presence of the NOS-II gene or of the targeting vector ²⁰ was accomplished by polymerase chain reaction (PCR) of DNA from tail biopsies. The animals were given food and water ad libitum, maintained under pathogen-free conditions of 12-hour light/12-hour darkness, and used for experiments at the age of 2 months. 

Corneal Cauterization

All experiments for this study were carried out according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The experiment was repeated three times; 16 wild-type and 16 knockout mice (6 weeks old) were deeply anesthetized with an intraperitoneal ketamine injection (200 mg/kg body weight). Novesine (oxybuprocaine chlorhydrate; Merck, Nogent sur Marne, France) eyedrops were applied to the right eye of each mouse, and the corneas of these eyes were then cauterized with a silver nitrate applicator (Argentinum Nitriticum; Braun, Melsungen, Germany) to induce the growth of new vessels. The applicator was held in place for 3 seconds, and excess silver nitrate was removed by rinsing the eyes with 10 ml of 0.9% NaCl solution and then gently blotting them with tissue paper. ^{21 22} The left, unburned, eyes served as controls. 

Visualization and Quantification of Neovascularization

Each time the experiment was run 4 animals were killed at 2, 4, and 7 days after corneal cauterization by a lethal injection of intraperitoneal pentobarbital. To fill the microvasculature and quantify corneal neovascularization, the upper body was perfused with NaCl 0.9% solution until the washed-out perfusion solution became clear. The mice were then further perfused with a mixture of 65% Araldite CY 223, 25% Hardener 2967 (a kind donation of Ciba, Ciba, Rueil-Malmaison, France), and 10% waterproof India ink (Pantone, Ashford, UK). The eyes were left in situ for 1 hour at 37°C to allow complete polymerization, then enucleated, and submerged in 4% phosphate-buffered neutral formaldehyde for at least 24 hours. ^{18 21} An Araldite polymer was chosen for neovascular visualization because of its greater homogeneity compared with gelatin mixtures used in equivalent rat models, and because of its consequent capacity for more complete filling of the vascular tree in the mouse. The cornea and a 1-mm rim of adjacent scleral tissue were then separated from each globe, and 3 full-thickness peripheral radial cuts made through the cornea to allow flattening. Corneal flatmount preparations were masked to minimize observer bias and then examined by computerized image analysis (Biocom, Ulis, France) using a modification of the method of Proia et al. ¹⁸ The average vessel length per flatmount was determined by measuring vessel length at every clock hour (12 measurements per cornea). 

Immunohistochemical Analysis

Eyes from wild-type and knockout mice were enucleated, immediately frozen in OCT (Tissue-Tek Sahuru, Bayer Diagnostics, Puteaux), and sectioned (10-μm thick). Sections from wild-type and knockout mice were collected on the same glass slides, and thus underwent the same immunohistochemical procedures. Before immunohistochemistry, they were fixed with 2% paraformaldehyde for 8 minutes at 4°C. Endogenous peroxidase activity was blocked with 1% H₂O₂ in 0.1% (wt/vol) Saponin (from Quillaja Bark/Sigma/St Quentin, France) in phosphate-buffered saline (PBS) for 60 minutes. Endogenous biotin was blocked with Avidin/Biotin Blocking Kit (Vector Laboratories, Burlingame, CA) in 5% (wt/vol) milk in PBS for 30 minutes. The slides were incubated overnight with the primary antibody diluted at 1:50 or 1:100 in 0.1% (wt/vol) Saponin–1% (wt/vol) milk in PBS at room temperature. The primary antibodies used were either polyclonal NOS-II antibody (Transduction Laboratory, Interchim, Montluçon, France), rat monoclonal anti-mouse F_4/80 antibody (Serotec; Argene, Varilhes, France), or rat monoclonal anti-mouse neutrophils antibody (Serotec). After washing, the sections were incubated in a 1:300 solution of biotinylated secondary antibody (goat anti-rabbit or rabbit anti-rat IgG) for 60 minutes, followed by further incubation with streptavidin–horseradish peroxidase for 45 minutes. The immunocomplex was revealed using 3.3′ diaminobenzidine tetrahydrochloride in the presence of H₂O₂. Finally, the sections were counterstained with Hemalun. Control experiments omitting the first antibody gave no staining (data not shown). 

RNA Isolation and Reverse Transcription–Polymerase Chain Reaction Analysis

Total RNA from the cornea was isolated at different times after cauterization by the acid guanidinium thiocyanate–phenol-chloroform method. One microgram of RNA was reverse-transcribed for 90 minutes at 42°C with 200 U of superscript Moloney Murine Leukemia virus reverse transcriptase (Life Technologies SARL; Eragny, France), using random hexamers, and 2 μl of cDNA was added to each polymerase chain reaction (PCR), as previously described. ²³ Amplification was performed as follows: 94°C for 2 minutes; 24 cycles for GAPDH and 30 cycles for NOS-II (number of cycles below the saturating conditions) at 94°C for 30 seconds, 55°C for 30 seconds, 72°C for 45 seconds; and then 72°C for 2 minutes. The amplified fragments were separated in a 1.2% agarose gel and transferred onto a nylon membrane (Amersham, Les Ulis, France). Specificity of the amplification process was verified by hybridization of blots with a ³²P-labeled specific internal oligonucleotide probe. The fragments were then washed 3 times in 1× SSC, 0.1% sodium dodecyl sulfate at 50°C. Visualization was achieved by exposure of x-ray film to the fragments. For intensity measurements the x-ray films were scanned, and the bands were quantified using densitometric measurements (NIH Image). The nucleotide sequences of the oligonucleotide primers used for reverse-transcription (RT)–PCR and those of hybridization probes are as follows: NOS-II antisense (TGTGTCTGCAGATGTGCTGAAAC); NOS-II sense (TTTCTCTTCAAAGTCAAATCCTACCA); NOS-II hybridization probe (GGGTCGATGTCACATGCAGCTTGTCCAGGGA); GAPDH antisense (ATGGCATGGACTGTGGTCAT); GAPDH sense (ATGCCCCCATGTTTGTGATG); and GAPDH hybridization probe (GCTGACAATCTTGAGGGAGTTGTCATATTT). 

Statistical Analysis

Results were expressed as mean ± SEM. Statistical analyses were performed using the nonparametric Mann–Whitney rank sum test, and P < 0.05 was considered significant. 

Results

To determine whether NOS-II is associated with the development of corneal alkali burns, RNA was extracted from the cornea at different times after cauterization, and NOS-II mRNA expression was then examined by RT–PCR analysis. The PCR product (amplified using NOS-II–specific primers) showed one specific band at the predicted size of 650 bp, which hybridized with the ³²P-labeled specific internal oligonucleotide probe (Fig. 1) . The NOS-II signal was absent in noncauterized wild-type control eyes. Expression of NOS-II mRNA was detected in the cornea at 2 days after cauterization, with maximal detection at 4 and 7 days after cauterization (Fig. 1A) . No expression was detected in cauterized or noncauterized knockout eyes (Fig. 1B) . 

Using immunohistochemistry, we attempted to localize the cells expressing NOS-II in the cornea 4 days after cauterization. The immunohistochemistry on wild-type sections revealed NOS-II expression in a subgroup of migratory cells in the cauterized cornea, which was maximal in the cauterized center of the cornea and decreased in density toward the limbus (Fig. 2 A). The limbal cornea itself did not show any NOS-II–positive cells. Some of the NOS-II–positive cells were also found in the adjacent granulomatous tissue posterior to the cauterization site. Cauterized knockout sections did not show any positive cells at any stage of the disease (Fig. 2B) . No cells expressing NOS-II were detected in noncauterized wild-type sections (data not shown). To clarify what type of cells were at the origin of the NOS-II expression, immunohistochemical experiments were performed on adjacent sections, using either a specific antibody against neutrophils or an F_4/80 antibody, a marker for mature macrophages. No neutrophils were observed in noncauterized corneal sections from either wild-type (Fig. 3A ) or knockout (Fig. 3D) mice. Four days after the cauterization, a large number of neutrophils could be observed in both wild-type and knockout mouse corneas (Figs. 3B and 3E , respectively). However, no F_4/80-positive cells were detected in the center of the cauterized corneas of either the wild-type or knockout mice (Figs. 3C and 3F , respectively), and only very few were detected in the peripheral cornea. Immunohistochemical evaluation at 7 days after cauterization revealed the same pattern of expression for NOS-II–positive cells, neutrophils, and F_4/80-positive cells in both types of eyes (data not shown). 

To determine whether the expression of NOS-II is involved in neovascularization after corneal burns, the length of new vessels, a reliable indicator of neoangiogenesis, was compared between wild-type and knockout mice (Fig. 4) . Quantification of the corneal vessels revealed no difference in limbal vascular loop height in noncauterized wild-type and knockout mice (Figs. 4A , 4B , and 5 ). Two days after cauterization, new vessels started to grow centrally from the dilated limbal loops in both wild-type and knockout mice (Figs. 4C , 4D ). At this stage of the disease, no differences in vessel length were detectable (Fig. 5) . The length of the new vessels greatly increased at 4 days after cauterization in both types of mouse (Figs. 4E and 4F ), with maximal growth 7 days after cauterization (Figs. 4G and 4H ). Measurement of the new vessels (Fig. 5) revealed that vessel length in wild-type mice was significantly greater than in knockout mice at 4 days after cauterization. This difference became more accentuated at 7 days after cauterization, suggesting that the lack of NO release promotes the growth of new vessels. 

Discussion

In this study, we demonstrated that NOS-II is expressed during the inflammatory reaction that follows severe alkali burns of the cornea in mice. RT–PCR assay showed that cauterized corneas expressed NOS-II mRNA. The mRNA appeared 2 days after cauterization and was maximal between 4 and 7 days after cauterization, thus corresponding to the severity of the disease (evaluated by the clinical appearance and the length of new corneal vessels). 

With regard to the origin of NOS-II in the burned cornea, immunohistochemical experiments revealed the presence of NOS-II in the central cornea and the adjacent granulomatous tissue, in which a maximum number of inflammatory cells was observed. Although the cells expressing NOS-II were not identified by double-labeling procedures, our immunohistochemical experiments gave some indications concerning the identity of NOS-II–positive cells. In the cornea, 4 days after cauterization, most of the migratory cells became stained with the specific anti-neutrophil antibody, whereas only a few cells could be detected with the specific antibody for mature macrophages (≈6/section). Our results match those of a previous report, ¹⁹ which shows that the infiltrating cells in this model of sterile inflammation consist mainly of neutrophils and monocytes and very few mature macrophages. All these data suggest that a subgroup of neutrophils or inflammatory monocytes may be the major source of NO in the cornea during the disease. Langerhans cells were observed in the central corneal epithelium after thermal cauterization. ²⁴ This type of cell might express NOS-II in our model, even though they have never been detected during chemical cauterization. ¹⁹ As we recently demonstrated, NOS-II can be expressed in vitro in corneal keratocyte after cytokine stimulation, ²⁵ and we therefore cannot exclude the possibility that these cells might also be able to contribute to the NO production in vivo during severe corneal alkali burns. 

Furthermore, using mice lacking the NOS-II gene, we demonstrated that production of NO by the NOS-II enzyme partially reduced the corneal angiogenesis during the disease. This stands in apparent contradiction to the results of Ziche et al., ^{9 17} who demonstrated a proangiogenic role of NO in the cornea. These authors showed an increase of corneal neovascularization by adding NO donors to a Cornea Pocket Assay with a pellet containing PGE₁ or SP and a decrease of PGE₁-, SP-, and VEGF-induced corneal neovascularization through the presence of NOS inhibitors. There are different possible explanations for the differences between our results and the results obtained with the Corneal Pocket Assay. First, the mechanisms involved in corneal neovascularization in the cauterized cornea are of a more complex nature than those involved in neovascularization induced by a unique compound in the Corneal Pocket Assay, making them difficult to compare. Second, it seems likely that the expression of distinct NOS isoforms differs from one model of corneal neovascularization to another, a difference that would have remained undetected in the experiments of Ziche et al. ^{9 17} because they used nonspecific NOS inhibitors. 

The mechanisms by which NO inhibits angiogenesis during corneal cauterization are as yet unknown. The neovascularization differences in our model can be explained either by a direct effect of NO on the angiogenesis process or by an indirect influence of NO on angiogenesis via a difference in the inflammatory reaction consecutive to cauterization. One might presume that the duration of the disease in the mouse expressing NOS-II is shorter than in the knockout mouse, because NO produced by NOS-II ought to decrease the amount of inflammatory cells by inducing their apoptosis. ^{26 27} However, we did not have the impression that there was a major difference in the inflammatory reaction between wild-type or knockout mice in terms of the infiltration by neutrophils and in the clinical presentation at any stage of the disease. 

Nitric oxide may possibly influence corneal angiogenesis by modulating the release or the expression of angiogenic factors present during severe corneal alkali burns. In this respect, VEGF and fibroblast growth factor have been described as important angiogenic factors involved in ocular ²⁸ and more specifically in corneal ^{29 30} neovascularization. Because NO is known to downregulate VEGF expression in vitro ³¹ and in vivo, ¹³ it is possible that a similar mechanism operates in burn-related corneal neovascularization. This possibility is currently under investigation in our laboratory. We cannot, moreover, exclude the possibility of a release of anti-angiogenic factors ^{32 33} (e.g., angiostatin, endostatin) triggered by NO, even though the involvement of these factors in corneal neovascularization has never been demonstrated. 

In our model, we observed an increasing inhibition of angiogenesis with time in the wild-type mice compared with the NOS-II–deficient mice. Because NOS-II expression is mainly found at the cauterization site (i.e., in the central cornea), we can assume that the angiogenic front is submitted to an increasing NO concentration as it grows toward the center. Through the addition of different NO donors, it has been shown that exogenous NO inhibits endothelial cell proliferation ^{34 35} and migration. ³⁶ Similar phenomena could account for the inhibitory effect of NO on new vessel formation in the cauterized mouse cornea. 

Our results conclusively show that NO released by NOS-II inhibits angiogenesis after severe corneal alkali burns. These findings may lead to a new therapeutic approach to treat intracorneal neovascularization. 

 Supported by Deutscher Akademischer Austauschdienst.

 Submitted for publication February 8, 1999; revised June 10, 1999; accepted June 28, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Kinetics of NOS-II mRNA expression in the cornea after cauterization. One microgram of total RNA extracted from the cornea of wild-type (A) or NOS-II knockout (B) mice at the day indicated after cauterization was used for each RT–PCR. PCR products were identified using specific probes as described in the Materials and Methods section. Data in (C) represent the relative amount of NOS-II mRNA normalized to the relative amount of GAPDH mRNA in wild-type corneas, and values are mean ± SEM. This experiment represents 1 of 3 independent RT–PCR analyses with different RNA extracts that gave similar results.

Figure 2.

 

Immunocytochemistry of NOS-II in corneas 4 days after cauterization. Sections of corneas 4 days after cauterization from wild-type (A and C) and knockout mice (B and D) were incubated with the NOS-II antibody as described in the Materials and Methods section. (A and B) General view of the cornea showing a gradient of NOS-II expressing cells in wild-type sections, which is highest in the central cornea (to the right) and decreases toward the limbus (to the left). (C and D) Magnification of the central cornea showing anti–NOS-II–positive cells scattered through the edematous stroma and the adjacent granulomatous tissue. The immunostaining shown is representative of three different experiments. Scale bar, 100 μm.

Figure 3.

 

Immunolocalization of neutrophils and monocytes/macrophages in corneas 4 days after cauterization from wild-type (A--C) and knockout (D–F) mice. Sections of noncauterized corneas (A and D) and sections of cauterized corneas (B, C, E, and F) were incubated with either anti-neutrophil antibody (A, B, D, and E) or anti-F_4/80 antibody (C and F) and counterstained with hemalun. The immunostaining shown is representative of three different experiments. The immunohistochemistry with the anti-neutrophil antibody demonstrated that there are no neutrophils in control uncauterized corneas in wild-type (A) or knockout (D) mice, whereas most of the infiltrating cells in cauterized corneas stained with the anti-neutrophil antibody in wild-type (B) as in knockout (E). No mature macrophages were found in the central cauterized cornea in either group of mice (wild-type in C; knockout in F), and only very few macrophages could be detected over the whole cornea (not shown). No difference in edema could be observed between wild-type and knockout mice at any stage of the disease. Scale bar, 100 μm.

Figure 4.

 

Flatmounts of corneas at different times after cauterization. View of corneal flatmounts from wild-type mice (A–D) and NOS-II knockout mice (E–H) at 0 (A and E), 2 (B and F), 4 (C and G), and 7 days (E and H) after cauterization. These photographs are representative of three different experiments. Scale bar, 500 μm.

Figure 5.

 

Quantification of corneal neovascularization at various times after cauterization in wild-type and knockout mice. Corneal neovascularization in wild-type (hatched bars) and in NOS-II knockout mice (solid bars) was evaluated after cauterization at the time indicated, as described in the Materials and Methods section. Results are mean ± SEM of three independent experiments. *P < 0.05 between wild-type and knockout mice.

The authors thank Laurent Jonet for technical assistance, Hervé Coet for photographic work, Jean Claude Jeanny and Holger Baatz for helpful discussion, and Andrew Ayers for critical review of the manuscript.