Netrin-1 Simultaneously Suppresses Corneal Inflammation and Neovascularization

Abstract

Purpose.: To investigate the effect of netrin-1 on alkali burn-induced corneal inflammation and neovascularization.

Methods.: The expression of netrin-1 and its receptors UNC5A, UNC5B, UNC5C, UNC5D, adenosine 2b receptor (A2BAR), deleted in colorectal cancer (DCC), and neogenin in normal and alkali-burned rat cornea were determined by RT-PCR and/or Western blot analysis, or immunostaining. Topical netrin-1 protein was applied to treat rat corneal alkali-burn injury for 14 consecutive days, started right after the injury or 10 days postinjury. Corneal inflammation and neovascularization were observed under slit lamp microscope. The apoptosis of corneal cells was determined by terminal deoxynucleotidyl transferase-mediated nick end labeling assay. Corneal inflammatory cell infiltration was evaluated by immunostaining of anti-PMN and anti-ED1 antibodies. The expression of epidermal growth factor (EGF), vascular epidermal growth factor (VEGF), and pigment epithelium-derived factor (PEDF) in rat cornea was determined by Western blot analysis.

Results.: Netrin-1 and its receptor UNC5B were expressed in normal rat corneal epithelium and stromal cells, and their expression decreased after corneal alkali burn. Exogenous netrin-1 administered on rat ocular surfaces resolved alkali burn-induced corneal inflammation, and also suppressed corneal neovascularization. Furthermore, netrin-1 could reverse neovascularization in alkali-burned cornea. The authors found that netrin-1 executed the functions through various mechanisms, including upregulating EGF expression, accelerating epithelial wound healing, inhibiting neutrophil and macrophage infiltration, reducing corneal cell apoptosis, and restoring the equilibrium of VEGF and PEDF in the wounded cornea.

Conclusions.: Netrin-1 could dampen inflammation, inhibit, and reverse neovascularization in alkali-burned cornea.

Netrin-1 is a diffusible, laminin-related protein that was first described as a guidance cue during neurogenesis. ^1,2 Canonical receptors for netrin-1 include deleted in colorectal cancer (DCC), the DCC paralogue neogenin, and the UNC5 homologs UNC5A, UNC5B, UNC5C, and UNC5D (for reviews, see Refs. 3, 4). Recently, the integrins α6β4, α3β1, ^5,6 and the adenosine 2b receptor (A2BAR) ^{7 –9} were also recognized as netrin-1 receptors. During development, netrin-1 is mainly expressed in the central nervous system. ² In adult mammals, netrin-1 and its receptors are also expressed in many nonneural tissues, such as the mammary glands, ¹⁰ pancreas, ⁵ lungs, ¹¹ kidneys, ¹² intestines, and liver and spleen. ¹³ Moreover, netrin-1 is also found to be highly expressed in human metastatic breast tumors ¹⁴ and aggressive neuroblastoma. ¹⁵ The broad expression of netrin-1 indicates its sophisticated function under both physiologic and pathologic conditions. 

Netrin-1 is known as an attractive or repellent guidance cue for axonal growth cones and neurons, depending on the receptors to which it binds. ^{16 –18} Attraction is mediated by the DCC receptor, ¹⁹ whereas repulsion requires receptors from the UNC5 family, acting alone or together with DCC family receptors. ²⁰ Other than its roles in axon pathfinding, netrin-1 has also been shown to impact mammary gland development, ¹⁰ modulate the morphogenesis of the lungs ¹¹ and vascular system, ^20,21 regulate the apoptosis of intestinal cells, ²² and promote tumorigenesis. ²³  

Recently, netrin-1 was found to be a negative guidance cue for leukocyte migration, ^8,12 which indicates the anti-inflammatory potential of netrin-1. Several in vivo studies have been conducted to evaluate the effect of netrin-1 on animal disease models, such as subcutaneous application for experimental colitis, ²⁴ intraperitoneal injection for peritonitis, ¹³ and inhalation or intravenous injection for lipopolysaccharide-induced pulmonary injury. ⁹ These studies have shown the potent effect of netrin-1 on reducing neutrophil infiltration, while the mechanism is variously dependent on receptor A2BAR ^9,13,24 or UNC5B. ¹²  

Furthermore, netrin-1 was found to be involved in the process of angiogenesis. However, there is dispute regarding the role of netrin-1 in neovascularization. ²⁵ Different studies have shown that netrin-1 is an angiogenic factor ^{21,26 –29} or an antiangiogenic factor. ^20,30,31 Others have shown netrin-1 exhibits a proangiogenic function in low concentrations and an antiangiogenic function in high concentrations. ³² Notably, previous studies used in vitro cell culture assay, ^26,27 aortic ring sprouting assay, ²⁰ chorioallantoic membrane assay, ^26,31 in vivo corneal micropocket assay, ²⁶ zebrafish development model, ²¹ or hind limb ischemia murine model ²¹ to study the effect of netrin-1. These models mimic different facets of the complicated procedure of blood vessel formation during morphogenesis, and tumorigenesis or inflammation, in which netrin-1 may play different roles. 

As is well known, inflammation and angiogenesis present simultaneously in many pathologic situations, and inflammation is one of the major reasons for angiogenesis in diseases such as arthritis, ischemic vascular diseases, and certain cancers (for reviews, see Refs. 33 –35). Thus, both inflammation and angiogenesis are targets in the treatment of these diseases. Previous studies investigated the effect of netrin-1 on either angiogenesis or inflammation separately in different in vitro or in vivo models, but there have been no studies to determine whether netrin-1 resolves inflammation and simultaneously induces or reduces angiogenesis in a disease that coupled inflammation with neovascularization. Therefore, we sought to use a corneal alkali burn model to address this issue. 

The corneal alkali burn model is a well-established severe ocular surface disease model that causes corneal epithelial defects, prominent corneal inflammation, corneal neovascularization, and reduced corneal transparency. ^36,37 It is widely used to study the mechanism and therapies of inflammation and angiogenesis, due to the easy local administration of medicine, well accessible position for observation, and the relatively immune-privileged status of the cornea. ^38,39 Our studies applied this model to determine the effect of netrin-1 on inflammatory and angiogenic diseases, and found that the topical application of netrin-1 on the cornea can suppress and reverse corneal neovascularization. Simultaneously, it reduces corneal inflammation. Moreover, we found netrin-1 may execute its different functions through various mechanisms. 

Materials and Methods

Rat Corneal Alkali Burn and Treatment

Wistar rats (180–220 g, 2 months old, male) were used in the study. Animal experiments were performed in accordance with the guidelines of the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research, and the animal experimental procedures were approved by the Experimental Animal Committee of Xiamen University. All rats were confirmed as being free of ocular diseases before the experiments. Rat corneal alkali burns were conducted as previously reported. ^36,40 In brief, the animals were anesthetized with intraperitoneal ketamine (60 mg/kg), and a filter paper disc (3 mm in diameter) incubated with 1 M NaOH for 60 seconds was then placed on the central cornea of the right eye for 30 seconds. The ocular surface was then rinsed with 30 mL of PBS. 

After the alkali burn injury, the animals were randomly divided into two groups of equal size. Rats in one group were topically administered with PBS (four times per day, 10 μL each), and the rats in the other group were topically applied with recombinant mouse netrin-1 (R&D Systems, Minneapolis, MN) using pipette (four times per day, 10 μL each, at the concentration of 5.0 μg/mL in PBS). The treatments were administered for 14 consecutive days. In some experiments, the treatments started from day 10 after the injury and lasted for another 14 days to observe the effect of netrin-1 on corneal neovascularization regression. Normal rats without alkali burn injuries were used as controls. After different durations, the animals were euthanized and their eyes were enucleated. The corneas were then dissected and stored at −80°C for histologic studies or used for RNA or protein extraction. 

Slit Lamp Microscopic Observation

The animals were examined under a slit lamp microscope every day after the alkali burns. The corneal images were taken by an experienced researcher (YS). Corneal epithelial defects were determined by 0.1% fluorescein sodium staining of the ocular surface and observation under cobalt blue light. The images were processed with image-processing software (Image Pro Plus V6.0; Media Cybernetics, Silver Spring, MD). Corneal neovascularization was quantified by calculating the wedge-shaped area (S) of vessel growth with the formula: S = C/12 × 3.1416 × [r ² − (r − I)²], ⁴¹ (S is the area, C is time, I is the radius from the center to the border of vessel growth, and r is the radius of the cornea). The inflammatory index was analyzed based on parameters including ciliary hyperemia, central corneal edema, and peripheral corneal edema as previously described. ⁴²  

Apoptosis Detection Assay

To detect apoptosis of the corneal cells, the corneas were embedded in OCT, cross-sectioned, and subjected to terminal deoxynucleotidyl transferase-mediated nick end labeling (TUNEL) staining (DeadEnd Fluorometric TUNEL system; Promega, Shanghai, China), according to the manufacturer's protocol. Cellular nuclei were stained with 4′-6-Diamidino-2-phenylindole (DAPI), and apoptotic cells were examined under a laser confocal microscope (Fluoview 1000, Olympus, Tokyo, Japan). The cellular nuclei and apoptotic cells were counted in three sections from each sample. 

Immunostaining

Cryosections of 4 μm were air-dried at room temperature for 30 minutes and fixed in acetone for 10 minutes at −20°C. After that, sections were rehydrated in PBS, and incubated in 0.2% Triton X-100 for 10 minutes After three rinses with PBS for 5 minutes each and preincubation with 2% BSA to block nonspecific staining, samples were incubated with anti-PMN (1:4000, Fitzgerald, Newmarket Suffolk, United Kingdom) or ED1 (1:200, AbD Serotec, Oxford, United Kingdom) antibodies for 16 hours at 4°C. After three washes with PBS for 15 minutes, samples were incubated with a FITC-conjugated secondary antibody (goat anti-rabbit and goat anti-mouse IgG at 1:100, Sigma, St. Louis, MO) for 1 hour. After three additional PBS washes, they were counterstained with propidium iodide (1:1000), then mounted with an anti-fade solution and examined under a laser confocal microscope. For immunohistochemical staining of netrin-1 and UNC5B, endogenous peroxidase activity was blocked by 0.6% hydrogen peroxide for 10 minutes. Nonspecific staining was blocked by 1% normal goat serum for 30 minutes. Sections were then incubated with rabbit anti-netrin-1 (1:20; Santa Cruz, Santa Cruz, CA) and goat anti-UNC5B (1:20; Santa Cruz) antibodies for 16 hours at 4°C. After three washes with PBS for 15 minutes, they were incubated with biotinylated goat anti-rabbit IgG (1:100) rabbit anti-goat IgG (1:100) for 30 minutes, followed by incubation with ABC reagent for 30 minutes. The reaction product was developed with DAB for 2 minutes. For negative control of the immunostaining, rabbit or goat primary antibody isotype control (Invitrogen, Carlsbad, CA) was applied during the immunostaining procedure. 

For analysis of integrated optical density (IOD) expression of positive immunostaining, images from immunostained (PMN and ED1 proteins) sections were processed using image-processing software (Image Pro Plus V6.0; Media Cybernetics, Bethesda, MD), as described in our previous report. ⁴³  

RNA Isolation and Reverse Transcription Polymerase Chain Reaction

For RT-PCR, total cellular RNA was extracted from the entire rat cornea, separated corneal epithelium and stroma, and rat brain tissue using a reagent (Trizol; Invitrogen), according to the manufacturer's instructions. Three micrograms of purified total RNA were used as a template to generate the first strand of cDNA using murine leukemia virus reverse transcriptase (Moloney; Fermentas, Shenzhen, China). PCR was performed using the following primer pairs: Netrin-1, 5′-GAGTCCATGGCCATCTACAAG-3′ and 5′-AAAGCCTGTGATTGCCACC-3′; UNC5A, 5′-GCCGGCTGATGATCCCTA-3′ and 5′-GCTGAGTCCAGTCCAGTCGTTAGCTG-3′; UNC5B, 5′-AGAAGGGGAAGGCCAGATT-3′ and 5′-AGATACCACTGTCCATCCGC-3′; UNC5C, 5′-TGGAGTGGCTCTCAAAGAAA-3′ and 5′-TGGTATCGATTTGCCTCTCC-3′; UNC5D, 5′-GTCAGTTCAGAGCATTGGAA-3′ and 5′-GTAAAGCAGCTCAGCTCAAGGTGGC-3′; Neogenin, 5′-ATGTGGTGCATTCCAAACAC-3′ and 5′-TTTACCAGCCAGCCAGAACC-3′; DCC, 5′-CCCAAGCTGGCTTTTGTACT-3′ and 5′-TGTCTGAACCTTCTGATGCTG; and A2BAR, 5′-TTCTGCACGGACTTTCACAG-3′ and 5′-CAAAATCTTCATGGTGGCCT. Amplification was carried out (iCycle; Takara, Dalian, China) (93°C for 3 minutes, 93°C for 30 seconds, 55°C for 30 seconds, 68°C for 30 seconds × 40, and 68°C for 7 minutes). 

Western Blot Assay

To detect protein expression in the corneas, the entire corneal tissue was carefully dissected and extracted with cold PIPA buffer and proteinase inhibitor cocktail (Merck, Darmstadt, Germany). In some experiments, the corneal epithelium was removed with a cell scraper, and the epithelium and stroma were extracted separately. Equal amounts of proteins extracted from lysates were subjected to electrophoresis on 10% sodium dodecyl sulfate-polyacrylamide gels and then electrophorectially transferred to PVDF membranes. After 30 minutes of blocking in 3% BSA, the blots were incubated with primary antibodies netrin-1 (1:200; Santa Cruz), UNC5B (1:200; Santa Cruz), vascular epidermal growth factor (VEGF; 1:200; Santa Cruz), PEDF (1:200; Santa Cruz), and β-actin (1:10,000; Bio-Rad, Hercules, CA) as a loading control. After three washes with Tris-buffered saline with 0.05% Tween 20 for 10 minutes each, the membranes were incubated with horseradish peroxidase (HRP) conjugated secondary antibodies and IgG (goat anti-rabbit, 1:5000; Bio-Rad; goat anti-mouse, 1:5000; Bio-Rad; and rabbit anti-goat, 1:5000; Dako, Shanghai, China) for 1 hour at room temperature. The results were visualized by enhanced chemiluminescence reagents and recorded on film. 

Statistical Analysis

Summary data were reported as mean ± SD. Group means were analyzed using the Student's t-test, where P < 0.05 was considered statistically significant. The statistical analysis was conducted with software (GraphPad Prism for Windows, version 5.00; GraphPad Software Inc., La Jolla, CA). 

Results

Expression of Netrin-1 and Its Receptors in Normal and Alkali-Burned Rat Cornea

Previous studies have demonstrated the tissue distribution of netrin-1 in organs such as the brain, lungs, heart, kidneys, intestines, liver, and spleen, ¹² while the expression of netrin-1 in the cornea remains unknown. We performed RT-PCR on the entire rat corneal lysate and found that netrin-1 and its receptor UNC5B were highly expressed in normal rat cornea (Fig. 1F), while receptors such as A2BAR, UNC5A, UNC5C, UNC5D, DCC, and neogenin were not expressed (Fig. 1E). We then separated the corneal epithelium and stroma, conducted RT-PCR, and found that netrin-1 and UNC5B mRNA were predominantly expressed in the corneal stroma (Fig. 1C). We further performed Western blot analysis to confirm the protein expression of these two genes. Interestingly, the majority of these two proteins resided in the corneal epithelium, but not in the corneal stroma (Fig. 1D). We then performed immunohistochemical staining and found that netrin-1 (Fig. 1A) and UNC5B (Fig. 1B) were expressed in both the corneal epithelium and stroma. Strong staining was detected in the membrane of the basal epithelial cells. 

Figure 1.

 

Expression of netrin-1 and its receptors in normal and alkali-burned rat cornea. Netrin-1 (A) and UNC5B (B) are expressed in the stroma and epithelia of the normal rat corneas. The strong staining was detected in the basal epithelia (arrowheads). (C) Netrin-1 and UNC5B mRNA are expressed in the normal rat corneal epithelia and stroma. The mRNA levels of both genes are higher in stromal tissue than in the epithelia. (D) Netrin-1 and UNC5B proteins are expressed in the normal rat corneal epithelia and stroma. Both protein levels are higher in the epithelia than in the stroma. (E) Netrin-1 receptors UNC5A, UNC5C, UNC5D, DCC, and neogenin expression were not detected in normal rat corneas or in corneas at different time points after the alkali burns. As a positive control, all these genes were highly expressed in rat brain tissue. (F) In the PBS group, netrin-1 gene expression decreased at day 1 after the alkali burns and returned to a normal level from day 3 to day 7. The receptor UNC5B's gene expression decreased from day 1 to day 7, while the receptor A2BAR's expression gradually increased from day 1 to day 7. In the netrin-1 treatment group, netrin-1 gene expression showed a gradual increase from day 1 to day 7, and there was only a mild decrease in the expression of UNC5B. There was only mild expression of A2BAR at day 1, and this diminished from day 3 to day 7.

We then investigated the gene expression of netrin-1 and its receptors in alkali-burned rat cornea. There was a mild decrease in netrin-1 gene expression at day 1 after the alkali burn, though it returned to a normal level from day 3 to day 7 (Fig. 1F). The receptor UNC5B gene expression decreased from day 1 to day 7 (Fig. 1F), while receptor A2BAR expression gradually increased from day 1 to day 7 (Fig. 1F). Other receptors, including UNC5A, UNC5C, UNC5D, DCC, and neogenin did not show obvious expression throughout the duration (Fig. 1E). When netrin-1 was applied to the rat corneas after the alkali burn, the netrin-1 gene expression showed a gradual increase from day 1 to day 7, and there was only a mild decrease of UNC5B gene expression (Fig. 1F). However, other receptors, such as A2BAR, UNC5A, UNC5C, UNC5D, DCC, and neogenin did not show obvious expression (data not shown except for A2BAR). Immunohistochemical staining further confirmed that netrin-1 and UNC5B expression was similar to that of normal cornea in alkali-burned cornea treated with exogenous netrin-1 for 7 days, while their expression was lower in the PBS treatment group (data not shown). 

Netrin-1 Promotes Inflammation Dissolution after Alkali Burns

Alkali burn can cause severe corneal inflammation. ⁴⁴ One day after the alkali burns, the central stroma of the rat cornea appeared opaque and edematous (Fig. 2A). In the PBS group, the central cornea maintained opaque appearance and there was scar formation on day 14 (Fig. 2A). However, there was only mild edema and no scar formation in corneas treated with netrin-1 for 14 days (Fig. 2A). The inflammatory index showed slight decrease from day 1 to day 14 in PBS group, while there was dramatic reduction in the netrin-1 group, and there was significant difference between the two groups at day 7 and day 14 (Fig. 2B). 

Figure 2.

 

Netrin-1 prevents corneal neovascularization after alkali burns. (A) One day after the alkali burns, the central corneal stroma of the animals appeared opaque and edematous in both groups. On day 14 after the injury, new blood vessels reached the central corneas in the PBS group, and there was a remarkable decrease in corneal transparency. In contrast, there was only slight new blood vessel formation in the limbal areas in the netrin-1 treatment group, and the corneas remained transparent at day 14. (B) The inflammatory index of the ocular surface declined from day 1 to day 14 in both groups. However, it was significantly lower in eyes treated with netrin-1 at day 7 and day 14 (*P < 0.05 and **P < 0.01). (C) New blood vessel formation area (NV area) in the PBS group increased from day 1 to day 7, and there was a mild decrease at day 14 after the alkali burns. In contrast, corneas treated with netrin-1 showed only a mild increase of NV area at day 7. There was a significant difference between the two groups at day 7 and day 14 (**P < 0.01). (D) The average new blood vessel length (NV length) increased from day 1 to day 7 and decreased at day 14 in the PBS group, while the NV length continued to be very short in the netrin-1 treatment group. There were significant differences between the two groups at day 7 and day 14 (**P < 0.01). (E) hematoxylin and eosin (H&E) staining showed prominent new blood vessel formation from the limbi to the central corneas in the PBS group at day 14, which was well-indicated by the red blood cells that remained in the blood vessels. However, the netrin-1 treatment group only showed a few blood vessels in the limbi and none in the peripheral and central corneas.

To investigate whether netrin-1 can dampen the corneal inflammation in the late stage of alkali burn, we started netrin-1 treatment from day 10 post alkali burn. We found that there was corneal edema and scar reduction in the PBS group from day 10 to day 24. However, the corneal edema and stromal scar was almost completely vanished at day 24 in the netrin-1 group (Fig. 3A). The inflammatory index showed gradual decrease in the PBS group, while it was further decreased in the netrin-1 group at days 17, 20, and 24 (Fig. 3B). Collectively, these data indicated that netrin-1 accelerates the resolution of corneal inflammation after alkali burns in both early stage and late stage. 

Figure 3.

 

Netrin-1 promotes the regression of corneal neovascularization after alkali burns. (A) Ten days after the injury, dense neovascularization reached the central cornea. In this experiment, netrin-1 treatment began on day 10. By day 24, the new blood vessels regressed from the central cornea to the peripheral cornea in the PBS group. In contrast, almost all the new blood vessels had regressed to the limbal area in the netrin-1 treatment group by day 24. (B) The inflammatory index continuously decreased from day 10 to day 24 in both groups, while the index was significantly lower in the netrin-1 treatment group at days 17, 20, and 24 (*P < 0.05). (C) The NV area gradually reduced from day 10 to day 24 in both groups. There was a dramatic decrease of NV area in the netrin-1 treatment group at day 20 and day 24, and there was a significant difference between the two groups (*P < 0.05 and **P < 0.01). (D) The average NV length was continuously reduced from day 10 to day 24 in both groups, and the length was shorter in the netrin-1 treatment group at days 17, 20, and 24 than in the other group (*P < 0.05 and **P < 0.01). (E) H&E staining was performed on the rat corneas 24 days after the alkali burns. In the PBS group, the corneal epithelia were undulant and there were new blood vessels remaining in the anterior stroma of the peripheral and central corneas. The cellularity was relatively higher in the anterior stroma than in the posterior stroma. In contrast, the netrin-1 treatment group showed smooth epithelia and homogeneous stroma without new blood vessels. The stromal cellularity was similar to those of normal corneas.

Netrin-1 Prevents Corneal Neovascularization after Alkali Burns

Alkali burn can also cause severe corneal neovascularization. ⁴⁴ We then evaluated the effect of netrin-1 on corneal neovascularization by comparing the neovascularization area between the netrin-1 treatment and PBS groups at different time points. The onset of peripheral neovascularization occurred on day 1 after the alkali burns in both groups (Fig. 2A). In the PBS group, there was an ingrowth of new blood vessels toward the peripheral corneas on day 4, which reached the central corneas on day 7. After that, there was a mild regression of new blood vessels on day 14 (Figs. 2C, 2D). In contrast, corneas treated with netrin-1 showed only a mild increase in new blood vessels and maintained at low levels through out the whole duration (Fig. 2A). The new blood vessels' areas and lengths were much lower than those that were treated with PBS at day 7 and day 14 (Figs. 2C, 2D). H&E staining showed that, 14 days after the alkali burns, there was prominent new blood vessel formation in the anterior part of the corneal stroma in the PBS group, which was well-indicated by the remaining red blood cells in the blood vessels. However, the netrin-1 treatment group showed only a few blood vessels in the limbus and none in the peripheral and central corneas (Fig. 2E). These results suggested that netrin-1 can inhibit rat corneal neovascularization induced by alkali burns. 

Netrin-1 Promotes the Regression of Corneal Neovascularization after Alkali Burns

To investigate whether netrin-1 can induce the regression of well-formed corneal neovascularization, we began netrin-1 treatment on day 10 after the alkali burns. In the rat alkali burn model, corneal neovascularization reached its acme 10 days after the injury (Fig. 2C). After that, there was gradual blood vessel regression (Fig. 3C), even in those eyes to which only PBS was applied. By day 24 after the injury, the blood vessels had regressed from the central cornea to the peripheral cornea in the PBS group. In contrast, there was a rapid regression of corneal neovascularization when the eyes were treated with netrin-1, and almost all the blood vessels had regressed to the limbal area in the netrin-1 group by day 24 (Fig. 3A). The corneal neovascularization area (Fig. 3C) and length (Fig. 3D) on day 24 were significantly lower in the netrin-1 group compared with those of the PBS group. H&E staining showed that, 24 days after the alkali burns, there was undulance of the corneal epithelia in the PBS group, new blood vessels were sporadically present in the anterior stroma of the peripheral and central corneas, and the cellularity was higher in the anterior stroma than the posterior stroma. In contrast, the netrin-1 group showed smooth epithelia, homogeneous stroma without new blood vessels, and the cellularity was similar to those of normal corneas (Fig. 3E). These data indicated that netrin-1 treatment could regress neovascularization resulting from alkali burns. 

Netrin-1 Promotes Corneal Epithelial Wound Healing after Alkali Burn

To investigate the mechanism that netrin-1 implements its effect on alkali burn-induced corneal inflammation and neovascularization, we first investigated the corneal epithelial wound healing after the alkali burns. The fluorescein staining showed that corneal epithelial defects were completely healed on day 7, if the corneas were treated with 5.0 μg/mL netrin-1, while the corneas treated with PBS did not heal (Fig. 4A). Epithelial defect quantification showed a significant difference between the two groups at day 7 (Fig. 4B). The results of Western blot analysis showed a dramatic increase in epidermal growth factor (EGF) expression in the netrin-1-treated corneas compared with those of the PBS group at day 7 (Fig. 4C). These results indicated that netrin-1 could promote early re-epithelialization after alkali burns, and this may occur through the upregulation of EGF expression in the cornea. 

Figure 4.

 

Netrin-1 promotes corneal epithelial wound healing after alkali burns. (A) Fluorescein staining showed central corneal epithelial defects at day 1 (D1) after the alkali burns, and there was gradual decrease of the epithelial defect areas at day 4 (D4) and day 7 (D7) in both groups. The epithelial defects were completely healed on day 7, if the corneas were treated with netrin-1, while the corneas treated with PBS did not heal. Images at different time points are sequential pictures from the same rat eye treated with either PBS or netrin-1. (B) Epithelial defect quantification showed a significant difference between the two groups at day 7 (**P < 0.01). (C) Western blot analysis showed that there was a dramatic increase in EGF expression in netrin-1 treated corneas compared with those of the PBS group at day 7.

Netrin-1 Reduces the Alkali Burn-Induced Apoptosis of Corneal Cells

Alkali burns could cause direct corneal epithelia damage, which triggers apoptosis of underlying stromal cells. Excessive apoptosis promotes infiltration of inflammatory cells that can cause further damage of the tissue (for review, see Ref. 45). To determine the effect of netrin-1 on alkali burn-induced apoptosis, we performed a TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling) assay on corneas treated with PBS or netrin-1. As expected, there were no apoptotic cells present in normal rat corneas (Fig. 5B). However, the majority of the cells in the central corneas were TUNEL-positive at day 1 after the alkali burns, and the positive cells gradually decreased from day 4 to day 7 in the PBS group. At day 7, there were still many apoptotic cells in the basal epithelia and endothelia. In contrast, when the corneas were treated with netrin-1, there were much fewer apoptotic cells from day 1 to day 7 compared with the PBS group, and the majority of the apoptotic cells resided in the basal epithelia and endothelia (Fig. 5A). Statistical analysis showed a significant difference of apoptotic cells between the two groups at different time points (Fig. 5C). 

Figure 5.

 

Netrin-1 reduces alkali burn-induced apoptosis of corneal cells. (A) A TUNEL assay showed that, in the PBS group, the majority of the cells in the central cornea went into apoptosis at day 1 after alkali burns and the apoptotic cells gradually decreased from day 4 to day 7. At day 7, there were still many apoptotic cells in the basal epithelia and endothelia. In contrast, when the corneas were treated with netrin-1, the apoptotic cells mainly resided in the basal epithelia and endothelia at day 1 and dramatically decreased at day 4. There were only sporadic apoptotic cells in the corneal endothelia at day 7. (B) There were no apoptotic cells present in normal rat corneas. (C) A statistical analysis of the apoptotic cells on days 1, 4, and 7 between the two groups showed significant differences at all time points (***P < 0.001).

Netrin-1 Inhibits Inflammatory Cell Infiltration after Corneal Alkali Burns

To further illustrate the mechanism that netrin-1 inhibits corneal inflammation, we conducted immunostaining of PMN and ED1 antibodies to detect neutrophil and macrophage infiltration after alkali burns. PMN staining showed that the majority of the neutrophils were located in the limbal stroma and the anterior part of the peripheral and central corneal stroma at day 5 (Fig. 6A). When netrin-1 was administered to the eyes, there were only a few neutrophils detected in the corneal stroma (Fig. 6A). The IOD analysis showed that neutrophil infiltration became prominent at day 1 after the alkali burn, and gradually decreased from day 3 to day 7, and there were significant differences between the two groups at both day 5 and day 7 (Fig. 6B). ED1 staining showed that macrophage infiltration gradually increased and achieved the most significant state at day 7 after the alkali burn in the PBS group (Fig. 7B). The majority of the infiltrated macrophages were also located in the anterior part of the corneal stroma at day 7 (Fig. 7A). However, macrophage infiltration maintained a very low level from day 1 to day 7 in the netrin-1 group, and there were very few macrophages in the corneal stroma at day 7 (Fig. 7A). The IOD analysis showed significant differences between the two groups at both day 5 and day 7 (Fig. 7B). We also investigated the neutrophil and macrophage infiltration during the neovascular regression from day 10 to day 24. There were very few neutrophils at day 10, and became undetectable at day 24 in both groups. Macrophage infiltration decreased at day 10 compared with that of day 7, and became negative at day 24 in both groups (data not shown). 

Figure 6.

 

Netrin-1 inhibits neutrophil infiltration after corneal alkali burns. (A) Immunostaining of PMN antibody on corneal sections from 5 days after the alkali burns showed prominent PMN positive cells in the limbal, peripheral, and central corneal stroma in the PBS group. However, there were only sporadic PMN-positive cells distributed in the corneal stroma in the netrin-1 treatment group. (B) IOD analysis on PMN expression on corneal sections from different time points after the alkali burns showed a significant difference between the two groups at day 5 and day 7 (*P < 0.05 and **P < 0.01).

Figure 7.

 

Netrin-1 inhibits macrophage infiltration after corneal alkali burns. (A) An immunostaining of ED1 was performed to detect macrophage infiltration. At 7 days after the alkali burns, ED1 positive macrophages were abundantly distributed in the limbal stroma as well as in the anterior part of the peripheral and central corneal stroma in the PBS group. However, there were very few ED1 positive cells presented in the limbi and corneal stroma in the netrin-1 treatment group. (B) The IOD analysis of ED1 expression showed a significant difference between the two groups at day 5 and day 7 (*P < 0.05 and **P < 0.01).

Effect of Netrin-1 on VEGF and PEDF Expression after Corneal Alkali Burns

It is well established that corneal neovascularization is tightly regulated by a dynamic, natural equilibrium between local proangiogenic factors and antiangiogenic molecules. ^{46 –49} Among these factors, VEGF and PEDF are the major counterparts. To investigate the mechanism through which netrin-1 inhibits and reverses corneal neovascularization after alkali burns, we conducted Western blot analysis on VEGF and PEDF. The results showed that VEGF was expressed at low levels in normal rat cornea, while there was a dramatic increase at day 14 after alkali burns in the PBS group. In the late-stage treatment experiment, there was decrease in VEGF at day 24. In the netrin-1 group, the expression of VEGF was lower than those of the PBS group at day 14 and day 24 (Figs. 8A, 8B). PEDF was expressed in normal corneas and was dramatically decreased at day 14 and 24 after alkali burns. However, there was restoration of PEDF in the netrin-1 treatment group, although its expression was still lower than that seen in normal corneas (Figs. 8A, 8C). Collectively, these data suggested that netrin-1 downregulates VEGF and upregulates PEDF in corneas after alkali burns. 

Figure 8.

 

Effect of netrin-1 on VEGF and PEDF expression after corneal alkali burns. (A) Western blot analysis results showed that VEGF was expressed at low level in normal rat cornea, while there was a dramatic increase at day 14 (D14) after alkali burns in the PBS group. In the late-stage treatment experiment, there was a decrease in VEGF at day 24 (D24). In the netrin-1 treatment group, there was a dramatic downregulation of VEGF at day 14 and day 24. PEDF was expressed in normal corneas and was dramatically decreased after alkali burns at days 14 and 24, while it was restored after netrin-1 treatment. Densitometry of protein expression compared with β-actin showed significant differences between the PBS group and the netrin-1 treatment group in (B) VEGF and (C) PEDF at days 14 and 24 (**P < 0.01).

Discussion

For the first time, our study evaluated the effects of netrin-1 using an in vivo acute wounding model, rat corneal alkali burns, which presents both severe inflammation and angiogenesis. We found that netrin-1 could reduce corneal inflammation, and meanwhile, inhibit and reverse neovascularization in corneal alkali burns. 

In the present study, we first investigated the expression of netrin-1 in the cornea. We found netrin-1 mRNA was predominantly expressed in the corneal stromal cells (Fig. 1C), while the majority of the netrin-1 protein was detected in the epithelium (Figs. 1A, 1D). Thus, we propose that netrin-1 is mainly produced by keratocytes and secreted into the extracellular matrix. Corneal epithelial cells may sequester netrin-1 via the receptor UNC5B, which was preferentially expressed in the corneal epithelium (Fig. 1D), supporting the contention that the target cell of netrin-1 in normal cornea may be the epithelium. Because netrin-1 can bind to integrin α6β4 and α3β1, ^5,6 which were highly expressed in the basal corneal epithelial cells, ^50,51 we presume that netrin-1 may modulate epithelial-mesenchymal interaction through binding with integrins, or integrins may act as coreceptors of UNC5B. A similar pattern was also found in mammary gland morphogenesis, whereby netrin-1 is expressed in prelumenal cells and its receptor is expressed in adjacent cap cells. ¹⁰ Interestingly, exogenous netrin-1 applied to rat corneas induced endogenous netrin-1 mRNA expression, and the UNC5B receptor's expression also increased (Fig. 1F). This positive feedback may also play a role in netrin-1 therapy on corneal alkali burns. 

We noticed that one of the netrin family receptors, A2BAR, was not expressed in normal cornea, but it was gradually increased after alkali burns in the PBS group. However, there was only mild expression at day 1 in the netrin-1 treatment group, and it became undetectable after that (Fig. 1F). Because A2BAR is mainly expressed in inflammatory cells, ⁸ we speculate that A2BAR is produced in the cornea by infiltrated inflammatory cells after alkali burns. The low level of expression of A2BAR mRNA in the netrin-1 treatment group also fits the fact that there is only mild inflammatory cell infiltration after netrin-1 treatment (Figs. 6, 7). 

Our study clearly demonstrated that netrin-1 could dampen alkali burn-induced inflammation. Consistent with other acute inflammation models in tissues such as the colon, ²⁴ lungs, ^8,9,12 or kidneys, ⁵² netrin-1 inhibits neutrophil infiltration in alkali-burned cornea after 5 days of treatment (Fig. 6). Moreover, netrin-1 strongly inhibits macrophage infiltration in the intermediate phase of the wounding (Fig. 7). Different studies have shown the presence of neutrophils within the corneal stroma as early as 6 hours after corneal injury, peaks 24–48 hours after injury, and begins decreasing thereafter. ⁵³ Macrophage infiltration after injury starts later than neutrophils and lasts longer. The infiltrated neutrophils exert their phagocytic functions to clear pathogens and cellular debris, and their presence appears to facilitate wound closure. ⁵³ Macrophages act in concert with neutrophils to phagocytose debris and invading pathogenic microorganisms and are a source of growth factors that promote resolution of inflammation as well as cell migration and proliferation for wound healing. On the other hand, neutrophils release into the injured tissue oxidative, hydrolytic, and pore-forming molecules which can damage host cells; macrophages also secrete abundant inflammatory cytokines, chemokines, and angiogenic factors which contribute to angiogenesis and scar formation of the wounded tissue. Therefore, exaggerated or constant influx and presence of neutrophil and macrophage is detrimental. ^54,55 In our study, neutrophil and macrophage infiltration was significantly reduced approximately 1 week post injury, which may have a major impact on the resolution of corneal inflammation after alkali burns. 

Our study also showed potent antiangiogenic effect of netrin-1 in the treatment of corneal alkali burns. Netrin-1 applied in different stages of the alkali burns not only inhibited, but also reversed corneal neovascularization. This is the first time showing that netrin-1 simultaneously attenuates inflammation and neovascularization in one disease model. As it was shown before, inflammatory and traumatic disorders can disrupt the balance between angiogenic and antiangiogenic factors in the cornea, and tilt the balance toward angiogenesis. ⁵⁶ Our results demonstrated that netrin-1 could restore the balance between VEGF and PEDF that could be disrupted by the alkali burn. 

The major source of VEGF in the cornea is invading macrophages. ^57,58 VEGF can further amplify inflammatory corneal neovascularization by recruiting more macrophages. ⁴⁹ In our study, VEGF was downregulated after netrin-1 treatment, supporting the notion that netrin-1 may affect corneal neovascularization through inhibiting macrophage infiltration. Interestingly, PEDF expression was restored in netrin-1-treated corneas. PEDF is a potent antiangiogenic factor found in retinoblastoma cells, retinal pigment epithelium, iris, and cornea. ⁵⁹ PEDF promotes endothelial cell apoptosis and also inhibits endothelial cell migration and tube formation. ⁶⁰ The effect of netrin-1 on the expression of VEGF and PEDF may represent the machinery underlying the inhibitory effect of netrin-1 on corneal neovascularization. Future study is needed to explore the mechanism via which netrin-1 regulates PEDF expression. 

In this study, we also found that netrin-1 can promote corneal epithelial wound healing after alkali burns (Figs. 4A, 4B). This function may be conducted through different mechanisms. Firstly, EGF expression was restored after netrin-1 treatment, while it was at a much lower level in the PBS group (Fig. 4C). Netrin-1 has been shown to regulate the function of the EGF-like protein Cripto-1 ¹⁰ and affect the morphogenesis and differentiation of the mouse mammary gland. ⁶¹ However, the exact mechanism through which netrin-1 upregulates EGF expression remains unclear at this time. Secondly, the interaction of netrin-1 with integrins α6β4 and α3β1 can activate epithelial cell adhesion and migration. ⁵ Hence, it accelerates corneal epithelial wound healing. Thirdly, netrin-1 could promote epithelial wound healing through preventing corneal epithelial cell apoptosis. In the early stages of alkali burns, the majority of the central corneal epithelial cells went into apoptosis. However, netrin-1 significantly reduced the number of apoptotic cells, not only in the epithelia, but also among keratocytes and endothelial cells (Fig. 3). Several studies have shown the effect of netrin-1 on the apoptosis signaling pathway ²⁹ and the antiapoptotic function of netrin-1 in tumorigenesis. ²³ It has been shown that the dependent receptor DCC and UNC5H could induce apoptosis in the absence of netrin-1, while this effect can be blocked by the presence of netrin-1. ⁶² In our study, there is downregulation of netrin-1 after alkali burns (Fig. 1F). This may trigger UNC5B-induced cell apoptosis. Exogenous netrin-1 thus inhibits UNC5B-induced apoptosis. On the other hand, endogenous netrin-1 expression also increased after netrin-1 treatment (Fig. 1F). This may enhance the antiapoptotic effect of netrin-1 and accelerate corneal wound healing as a result. Furthermore, reduction of corneal cell apoptosis may also contribute to the decrease of macrophage infiltration, as a result, inhibits inflammation and neovascularization. 

Recent studies have shown the average plasma netrin-1 levels in a combined noncancerous diseases population were 479 ± 74 pg/mL, and there was a significant increase of netrin-1 in various cancers. ⁶³ The physiologic concentrations of netrin-1 in chicken brains ranged from 50 ng to 150 ng/mL, ^2,64 but the range is not known for other tissues. Previous studies have shown that netrin-1 stimulates angiogenesis at a low concentration of approximately 50 ng/mL, ²⁶ while it inhibits angiogenesis at high concentrations of approximately 1 μg/mL. ²⁰ In our study, we used netrin-1 at 5.0 μg/mL, a concentration higher than in other in vitro studies, on the treatment of alkali burns. The daily dose was 0.2 μg for each rat. Because we did not construct a dose curve in the present study, we cannot rule out that the underlying mechanism of antiangiogenesis is due to a high concentration of netrin-1. 

In summary, our study clearly demonstrated that exogenous netrin-1 application to the ocular surface could dampen inflammation, inhibit and reverse neovascularization, and accelerate epithelial wound healing of alkali burn-induced corneal damage. The effects of netrin-1 on the entire orchestration of the disease were executed through different mechanisms. Corneal neovascularization is one of the major reasons for corneal blindness and was a risk factor for rejection after allograft corneal transplantation. The multifunction features of netrin-1 may shed new light on the treatment of inflammatory and angiogenic diseases of ocular surface as well as other organs. 

Footnotes

 Supported by grants from the National Basic Research Program of China (2011CB504606 [ZuL]), the National Natural Science Foundation of China (NSFC No. 30872810 and 30931160432 [ZuL], 30872809 [WL], and 81160118 [YS]), the Technological Innovation Platform Program of Fujian Province (2009J1013 [ZuL]), the Natural Science Foundation of Fujian (No. 2009J06023 [WL]), Science and Technology Foundation of Jiangxi (No. 20111BBG70026-2 [YS]), and the Natural Science Foundation of Jiangxi (No. 20114BAB215029 [YS]).

Footnotes

 Disclosure: Y. Han, None; Y. Shao, None; Z. Lin, None; Y.-L. Qu, None; H. Wang, None; Y. Zhou, None; W. Chen, None; Y. Chen, None; W.-L. Chen, None; F.-R. Hu, None; W. Li, None; Z. Liu, None

The authors thank Hui He and Yuehong Ma for their technical support in the experiments.