Netrin-1 promotes diabetic corneal wound healing through molecular mechanisms mediated via the adenosine 2B receptor

Netrins are secreted chemoattractants with the roles in axon guidance, cell migration and epithelial plasticity. In the present study, we investigated the roles of netrin-1 in the regulation of corneal epithelial wound healing, inflammation response and nerve fiber regeneration in diabetic mice and cultured corneal epithelial cells. In diabetic mice, the expression of netrin-1 was decreased when compared with that of normal mice. Furthermore, high glucose blocked the wounding-induced up-regulation of netrin-1 expression in corneal epithelial cells. Exogenous netrin-1 promoted the corneal epithelial wound healing in diabetic mice, and facilitated the proliferation and migration by reactivating the phosphorylation of ERK and EGFR in high-glucose treated corneal epithelial cells. Moreover, netrin-1 decreased the neutrophil infiltration and promoted M2 macrophage transition, accompanied with the attenuated expression of pro-inflammatory factors in diabetic mouse corneal epithelium. The promotions of netrin-1 on corneal epithelial wound healing and inflammation resolution were mediated at least through the adenosine 2B receptor. In addition, netrin-1 promoted the regeneration of corneal nerve fibers that was impaired in diabetic mice. Taken together, netrin-1 regulates corneal epithelial wound healing, inflammation response and nerve fiber regeneration in diabetic mice, indicating the potential application for the therapy of diabetic keratopathy.

different signaling pathways and play distinct roles in physiological and pathophysiological conditions 13 . Other than the function of axon guidance, netrin-1 can also protect the kidney against ischemia-reperfusion injury, regulate angiogenesis, and even promote tumor growth [14][15][16][17] . In addition, netrin-1 was found recently to possess an anti-inflammatory capacity in lung injury and inflammatory peritonitis 18,19 . In cornea, a previous study has confirmed that netrin-1 can dampen alkali burn-induced inflammation and neovascularization 20 .
Based on the multiple functions of netrin-1 and the characteristics of diabetic corneal pathogenesis, we hypothesized that netrin-1 may assume a potential use for the treatment of diabetic keratopathy. To address this hypothesis, we studied the regulation and mechanism of netrin-1 on corneal epithelial wound healing, inflammatory response, and nerve fiber regeneration by using type 1 diabetic mice and high glucose-treated corneal epithelial cells.

Results
Hyperglycemia causes decreased netrin-1 expression in corneal epithelium. To examine the expression of netrin-1, mouse cornea and epithelium were collected and analyzed by using immunofluorescence staining, RT-qPCR and ELISA. The immunofluorescence staining of netrin-1 in normal and diabetic unwound mice corneal sections showed that the expression of netrin-1 in diabetic mouse was decreased (Fig. 1A). In unwounded diabetic mice, the mRNA transcripts of netrin-1 were significantly downregulated (Fig. 1B) when compared with that of control mice. Moreover, the netrin-1 concentration was reduced in diabetic mice corneal epithelium compared to that of control mice (Fig. 1C). Furthermore, the mRNA and protein levels of netrin-1 were downregulated in regenerating corneal epithelium of diabetic mice compared to those levels in control mice (Ctrl vs. Diabetic in Fig. 1 B,C). The results suggest that the netrin-1 expression was decreased and the function of netrin-1 upregulated stressed by wounding was impaired in diabetic mice.

Netrin-1 promotes migration, proliferation, and ERK and EGFR phosphorylation of corneal epithelial cells impaired by high glucose.
To determine the impact of netrin-1 on corneal epithelial wound healing in vitro, mouse corneal epithelial cells were incubated with high glucose, with an equal concentration of mannose as the osmotic control, until monolayer formation. To detect the migration of TKE-2 cell incubated with high glucose, the corneal epithelial monolayer was wounded and incubated with or without netrin-1 (50 ng/ mL) for 24 h. The results showed that corneal epithelial wound closure was significantly delayed by high glucose Figure 1. Hyperglycemia downregulates netrin-1 expression in corneal epithelium. Normal and diabetic mouse corneal epithelium was collected before (as unwound group) or 48 h after epithelial scrape (as wound group). Netrin-1 expression was examined by using immunofluorescence staining (A, 3 mice per group), RT-qPCR (B, 9 mice per group) and ELISA (C, 9 mice per group) with the age-matched normal mice as control. *p < 0.05, **p < 0.01.
SCIEntIfIC RepoRts | (2018) 8:5994 | DOI:10.1038/s41598-018-24506-9 treatment while netrin-1 promoted the migration capacity of high glucose-treated cells to the same level of normal cells (Fig. 3A,B). To evaluate the proliferation of netrin-1, we inoculated and cultured 10 5 TKE-2 cells with normal and high glucose medium with/without netrin-1. After 3 days, we collected and amounted the cells, the results showed that netrin-1 improved the proliferation potential of corneal epithelial cells that was impaired by high glucose treatment (Fig. 3C). To elucidate the mechanism underlying the promotion of netrin-1 on corneal epithelial wound healing, we investigated the effects of netrin-1 on the upregulation of EGF expression and activation of ERK and EGFR that were impaired in diabetic corneal epithelium. The immunofluorescence staining of EGF in mice corneal epithelium showed that the immunofluorescence luminance of EGF in diabetic group were weaker than normal group and netrin-1 treatment can effectively increase the expression of EGF in diabetic mice ( Figure S2). The results of western blotting showed that high glucose-cultured cells exhibited reduced activation of ERK and EGFR while netrin-1 significantly increased the phosphorylation levels of ERK and EGFR to even above that of control cells (Fig. 3D), which was also confirmed by immunofluorescence staining in the netrin-1-injected diabetic mouse corneal epithelium 48 h after injury (Fig. 3E).

Netrin-1 promotes inflammation regression in diabetic cornea. Persistent low-grade inflamma-
tion is known to be a large cause of diabetic complications. To explore the possible anti-inflammatory effects of netrin-1, recombinant netrin was used to treat diabetic, wounded corneas and its effects on the infiltrations of neutrophils, macrophages were examined at 48 and 72 h after the corneal epithelial scrape. Representative staining and flow cytometry were shown in Fig. 4. Netrin-1 decreased the infiltration of neutrophils (Ly6G+, 48 h, Fig. 4A,B) and increased the number of M2 macrophages (CD68+/ F4/80+CD206+, 72 h, Fig. 4A,C) in diabetic cornea compared to that of diabetic cornea without netrin-1 treatment. Moreover, the mRNA levels of iNOS and IL-12 (markers of M1 macrophages) were significantly decreased with the treatment of netrin-1 in diabetic cornea at 48 h after the corneal epithelial scrape, while the expressions of arginase-1 and IL-10 (markers of M2 macrophages) were increased with the netrin-1 treatment at 72 h after the corneal epithelium scrape (Fig. 4D). These results suggest exogenous netrin-1 promotes the resolution of inflammation during diabetic corneal wound healing.
Netrin-1 promotes diabetic corneal epithelial wound healing through Adenosine 2B receptor. In our preliminary experiment, we found that Adenosine 2B receptor and UNC5B receptor were significantly upregulated when cultured mouse corneal epithelial cells were either wounded or treated with netrin-1 ( Figure S3). To further determine whether A2BAR and/or UNC5B receptors mediate the improvement of netrin-1 on diabetic corneal wound healing, the A2BAR receptor-specific antagonist PSB1115 or anti-UNC5B blocking antibody was injected subconjunctivally in netrin-1-treated diabetic mice. The results showed that the A2BAR antagonist significantly attenuated the promotion of netrin-1 on diabetic corneal epithelial wound healing, with 20.77%±2.35% epithelial defect in antagonist-injected netrin-1-treated mice compared to only 11.59%±1.55% Figure 3. Netrin-1 promotes the migration, proliferation, and ERK and EGFR phosphorylation of corneal epithelial cells impaired by hyperglycemia. Confluent mouse corneal epithelial cells were wounded after incubation with 5 mM glucose, 30 mM mannose or 30 mM glucose. Cell migration was observed with or without 50 ng/mL netrin-1 treatment for 24 h (A). Migration rate was analyzed and represented as the percentage of primary wounding area (B, n = 3 per group). Cell proliferation was measured by cell counting after 3 days of netrin-1 treatment (C, n = 3 per group). The phosphorylated levels of ERK and EGFR were analyzed by Western blotting after 3 days incubation until monolayer formation. We administrated netrin-1 (50ng/ml) and after 8 or 15 minutes we collected the cells and then detected the p-ERK or p-EGFR expression by western blotting. (D, n = 4 per group). The expression of p-ERK/p-EGFR were also confirmed by immunofluorescence staining in the netrin-1 injected diabetic mouse corneal epithelium 48 h after injury (E, 3 mice per group). *p < 0.05, **p < 0.01, ***p < 0.001. defect in mice treated with netrin-1 alone and to 24.44%±4.09% defect in untreated diabetic mice at 72 h (Fig. 5A). Moreover, upregulation of the phosphorylated ERK level by netrin-1 was also blocked by the pretreatment of A2BAR antagonist in diabetic corneal epithelium, while the phosphorylated EGFR level was not changed with the treatment of A2BAR antagonist (Fig. 5B). However, the injection of UNC5B blocking antibody showed no effect on wound healing or the levels of phosphorylated ERK and EGFR in the corneal epithelium of diabetic mice (Fig. 5B). In addition, treatment with A2BAR antagonist but not with UNC5B blocking antibody abolished the regulation of netrin-1 on the infiltration of neutrophils and the phenotype transition of macrophages in  diabetic corneal wound healing (Fig. 5C,D). The results suggest that the adenosine 2B receptor is involved in the mediation of netrin-1 on the promotion of diabetic corneal wound healing.

Netrin-1 enhances nerve fibers regeneration in diabetic cornea.
To elucidate whether netrin-1 enhances corneal nerve fiber regeneration in diabetic mice, 50ng netrin-1(5 μL/eye) or PBS was injected twice a week subconjunctivally after the scrape of central corneal epithelium. After 14 days, the corneas were collected and whole-mount stained with β3-tubulin to mark the nerve fibers. The corneal subbasal nerve fibers and intraepithelial nerves 21 density in diabetic mice showed apparent decline compared with control mice, whereas mouse corneas with netrin-1 injection exhibited relatively densely in the subbasal nerve fibers and intraepithelial nerves (Fig. 6A,B,C). According to the analysis of present area occupied by subbasal nerves and intraepithelial nerves density, there was a significant difference in diabetic and netrin-1 injection mice (Fig. 6D,E)

Discussion
As one of the most densely innervated tissues, the cornea contains abundant nerve fibers that branch from the ophthalmic trigeminal ganglion. A number of neurotrophic factors and neuropeptide have previously been detected in the corneal epithelium, among which the pigment epithelium-derived factor (PEDF) 1 , nerve growth factor (NGF), glial cell-derived neurotrophic factor (GDNF) 22 and substance P 2 were involved in the regulation of corneal innervation in the corneal wound healing. As a traditional guidance cue for axon growth in the central and peripheral nervous system, previous study has confirmed the expression of netrin-1 in adult mouse corneal epithelium 20 . Herein this study, we found that netrin-1 expression was significantly decreased in both mRNA and protein levels in the corneal epithelium of diabetic mice when compared with normal mice. Furthermore, high glucose blocked the upregulation of netrin-1 expression induced by wounding in cultured corneal epithelial cells. Exogenous netrin-1 promoted corneal epithelial wound healing of normal and diabetic mice, and the proliferation and migration of high glucose-treated corneal epithelial cells.
Netrin-1 not only functions as an axon guidance factor in nervous system but also has many functions in non-nervous system, such as regulates the adhesion and migration of a variety of epithelial cells, including pancreatic epithelial cells 10 , mammary epithelial cells 23 , and renal proximal tubular epithelial cells 24 . In diabetic keratopathy, one of the most known characteristics is delayed corneal re-epithelialization caused by the impaired activation of the EGFR-, ERK-, and PI3K/AKT-signaling pathways 25 . Here, we demonstrated that exogenous netrin-1 facilitated diabetic corneal epithelial wound healing by upregulated of EGF expression and afterwards reactivating the phosphorylation of EGFR and ERK to promote the migration and proliferation of high glucose-treated corneal epithelial cells.
A2BAR and UNC5B are two classical receptors of netrin-1. Previous studies have shown that the potent effect of netrin-1 on inhibition of inflammatory response by reducing neutrophil infiltration in some non-ocular tissues is variously dependent on receptor A2BAR 18,19 or UNC5B 26 Moreover, A2BAR and UNC5B were the most two significant-decreased receptors in the expression level of diabetic corneal epithelium than that of control epithelium in vivo and significantly up-regulated after the mouse corneal epithelial cells were wounded in normal conditions. Combined with our preliminary results and previous descriptions in cornea, we subsequently investigated their roles in the promotion of netrin-1 in diabetic corneal wound healing by using the antagonist or blocking antibody. The in vivo application showed the A2BAR antagonist, but not the UNC5B blocking antibody, reversed the promotion of netrin-1 in diabetic corneal wound healing, which suggests that at least the A2BAR mediates the promotion of netrin-1 on diabetic corneal epithelial wound healing. However, considering to the possible different diffusing capacity caused by different molecular weight between antagonist and blocking antibody, the exact mechanism of netrin-1 on the promotion of diabetic corneal wound healing needs to be further investigated, including the possible involvement of other reported receptors 27 .
In acute normal wound healing, inflammation protects against infection and injury but must be timely resolved to prevent unwarranted tissue damage. The resolution of inflammation is accomplished by blunting neutrophil infiltration, decreasing pro-inflammatory factor production, promoting the macrophage phagocytosis of apoptotic cells and microbes, and stimulating the clearance of phagocytes to enable a return to homeostasis 23 . Moreover, the excessive inflammation and impaired inflammation resolution, characterized by excessive infiltration of neutrophils and transition impairment of M1-to M2-macrophage phenotype, has been demonstrated to contribute to impaired wound healing in diabetic mellitus 24 . In the present study, netrin-1 attenuated the infiltration of neutrophils and decreased the expression levels of pro-inflammatory factors in corneal wound healing of diabetic mice. In addition, netrin-1 also promoted the generation of M2-macrophages, which was impaired in the corneal wound healing of diabetic mice. The results were consistent with the previous descriptions of the anti-inflammatory activity of netrin-1 in diabetic nephropathy [28][29][30][31] , acute colitis 28,29,31 , pancreatitis 30 , lung injury 18 , and cornea by inhibiting neutrophil infiltration and inducing the M2 polarization of macrophages 20,32 . More interesting, when treated with A2BAR antagonist, rather than UNC5B blocking antibody, the induced the M2 polarization of macrophages was reversed, suggesting that the promotion of netrin-1 on M2 macrophage transition was signaling through the A2BAR, but not UNC5B.
The decreased corneal sensation and corneal nerve fiber density are the typical characteristics of diabetic keratopathy 33,34 . The dysfunction of corneal innervation is critical for the impairment of corneal re-epithelialization wound healing in diabetic mellitus 35 . In the present study, topical netrin-1 promoted the regeneration of diabetic mouse corneal nerve fibers in vivo. These findings suggest that netrin-1 expressed in corneal epithelium maintains the homeostasis of corneal epithelial innervation while the reduction of netrin-1 in corneal epithelium caused by hyperglycemia may be involved in the regression of corneal nerve fibers that is found in diabetic mellitus.
In summary, we show that netrin-1 regulates corneal epithelial wound healing, inflammation response, and nerve fiber regeneration in diabetic mice. The multiple function of netrin-1 provides a candidate for treatment of persistent corneal epithelial defects in patients with diabetic mellitus.  injection, with the blood glucose level above 350 mg/dL. We have used at least 114 normal mice and 300 diabetic mice in the experiments we showed.
Corneal epithelial wound healing. Mouse entire corneal epithelium, including the limbal region, was marked with 3 mm trephine and subsequently scraped with a corneal rust ring remover (Alger Co, Lago Vista, TX) in anesthetized normal and diabetic mice. After the corneal epithelium scraped, diabetic mice were injected subconjunctivally with 50 ng netrin-1 (5 μL/eye, R&D Systems, Minneapolis, MN) or equal phosphate buffer saline (PBS) as the vehicle control. For topical netrin-1 application, netrin-1 (50 ng in 2.5 μL PBS) was dropped onto the corneal surface by a 10 μL tip four times daily per eye for 3 days. For netrin-1 receptor inhibition, A2BAR antagonist PSB1115 (2.5 μg in 5 μL PBS, R&D Systems) or UNC5B receptor-blocking antibody (5 μg in 5 μL PBS, R&D Systems) was injected subconjunctivally at 24 h before and 0, 24, and 48 h after the scrape of corneal epithelium. Ofloxacin eye drops were applied to all mice to avoid infection. After 24, 48, and 72 h, corneal epithelium defects were visualized by staining with fluorescein sodium and photographed under slit lamp microscope (BQ900; Haag-Streit, Bern, Switzerland). The stained area on the residual epithelial defect was analyzed by using Image J software (National Institutes of Health, Bethesda, MD) with the original epithelial defect area as 100%.
Cell culture and treatment. The mouse corneal epithelial cell line TKE2, provided by Tetsuya Kawakita of Keio University (Tokyo, Japan), was cultured in keratinocyte serum-free medium (KSFM; Invitrogen, Carlsbad, CA) 36 . For the analysis of ERK and EGFR phosphorylation, the cells were starved overnight in bovine pituitary extract (BPE)-free KSFM (for ERK analysis) or BPE and EGF-free KSFM (for EGFR analysis) and subsequently incubated with 30 mM glucose or mannose (as osmotic control) in the presence of 10~100 ng/mL netrin-1.

Cell migration and proliferation analysis.
For migration analysis, the cells were incubated with high glucose until confluence, wounded with a micropipette tip, and incubated with or without 50 ng/mL netrin-1 for 24 h. Cells were photographed, and the wound closure was measured by Image J software. To determine the effect of netrin-1 on the proliferation of corneal epithelial cells, the TKE2 cells were starved overnight in BPE-free KSFM and subsequently incubated with 20 ng/mL netrin-1 for 3 days. The cell number was measured with an automated cell counter (Invitrogen Life Technologies, Paisley, UK).
Immunofluorescence and corneal whole mount staining. Frozen sections (8 μm) were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and blocked with 10% bovine serum albumin at room temperature. The samples were incubated with primary antibodies overnight at 4 °C and subsequently with corresponding secondary antibodies incubate for 1 h at room temperature. The information for primary and secondary antibodies are showed in Table 2. All samples were counterstained with 4' ,6-diamidino-2-phenylindole (DAPI, Sigma), and images were captured with an Eclipse TE2000-U microscope (Nikon, Tokyo, Japan). Corneal whole-mount immunofluorescence staining was performed as we and other researchers described 21,37,38 . Mouse eyeballs were collected and fixed in Zamboni's fixative for 1 h, then the cornea was dissected around the scleral-limbal region and blocked by PBS with 0.1% Triton X-100, 2% goat serum, and 2% bovine serum albumin for 2 h, and subsequently incubated in the same incubation buffer with Alexa Fluor ® 488 conjugated neuronal class III β-tubulin mouse monoclonal antibody overnight at 4 °C. After washing for 5 times, the flat mounts were examined under an Eclipse TE2000-U microscope. The quantification of corneal innervation was calculated as the percentage of area positive for β-tubulin staining as previously described 39-41 . Reverse-transcription quantitative polymerase chain reaction. For reverse-transcription (RT) quantitative polymerase chain reaction (qPCR), total RNA was extracted from the whole cornea (for detection of pro-inflammatory factor expression, 2 corneas were pooled as a sample) or corneal epithelium (for detection of netrin-1 and receptors expression, 2 corneas were pooled as a sample) or cultured TKE2 cells, using Nucleospin RNA kits (Macherey-Nagel, Düren, Germany) according to the instructions supplied by manufacturer. Corneal epithelium was collected by scraper under the microscope after the mice were executed. Western blotting analysis. Total protein was extracted from lysed samples of the mouse corneal epithelium (3 corneas were pooled as a sample) or cultured TKE2 cells and run on 10% SDS-PAGE gels before the transfer to a PVDF membrane (Millipore, Billerica, MA). The samples were blocked with 5% nonfat dry milk and incubated with primary antibodies overnight at 4 °C and subsequently with horseradish peroxidase-conjugated secondary antibody (Amersham Biosciences, Piscataway, NJ) for 1 h at room temperature. The information for primary antibodies are showed in Table 2. Finally, the blots were visualized via enzyme-linked chemiluminescence by using the ECL kit (Pierce Biotechnology, Rockford, IL) and quantified by using Image J software.
Enzyme-linked immunosorbent assay. Total protein was extracted from the intact or regenerating corneal epithelium (48 h after the scrape) of three eyes of normal or diabetic mice. All samples were centrifuged, and the supernatants were subjected to quantitative sandwich immunoassay with the enzyme-linked immunosorbent assay (ELISA) detection kit for mouse netrin-1 (USCN Life Science, Wuhan, China) according to the manufacturer's instructions. Absorbance was measured with a microplate reader (Molecular Devices, Sunnyvale, CA) at a wavelength of 450 nm.
Flow cytometry. Mouse eyeballs were collected and fixed in PBS at 48 h or 72 h after injury. Cornea was dissected around the scleral-limbal region and cut into pieces by corneal scissors. We incubated them in 50 uL Liberase TL (2.5 mg/ml; Sigma) for 30 min at 37 °C with gentle stirring and then undigested tissue removed by screen filters (4 corneas were pooled as a sample). We finally collected cell suspensions and incubated them with fluorescein-conjugated anti-mouse antibodies for 30 min at 4 °C and analyzed by flow cytometry with a Vantage flow cytometer (Becton Dickinson) and FlowJo (Tree Star, Inc.) software. The gate was set on CD45+ population, and further analysis of surface markers was done with in this gate. The primary antibodies used were as follows: CD45-PE/-cy5, CD86-PE-cy5, F4/80-FITC, CD11b-FITC, Ly6G-PE and CD206-PE (Biolegend, San Diego, CA). All experiments were repeated at least 3 times.
Statistical analysis. Data in this study were representative of at least three different experiments and were presented as the means ± standard deviation. Statistical analysis was performed using SPSS 17.0 software (SPSS, Chicago, IL) and one-way analysis of variance. Differences were considered statistical significance at p < 0.05.