DNase I improves corneal epithelial and nerve regeneration in diabetic mice.

DNase I has been reported to improve diabetic wound healing through the clearance of neutrophils extracellular traps (NETs) caused by neutrophil aggregation. However, the function of DNase I on diabetic corneal wound healing remains unclear. Here, we investigated the effect and mechanism of topical DNase I application on diabetic mouse corneal epithelial and nerve regeneration. Corneal epithelial defects, inflammatory response, regeneration-related signalling pathways, oxidative stress, corneal innervation and sensation were examined and compared between the diabetic and normal mice. The results confirmed firstly the increased NETs production during the delayed corneal epithelial wound healing of diabetic mice, which was significantly improved through either DNase I or Cl-amidine administration. Mechanistically, DNase I improved inflammation resolution, reactivated epithelial regeneration-related signalling pathways and attenuated the accumulation of reactive oxygen species (ROS). Moreover, DNase I application also promoted corneal nerve regeneration and restored the impaired corneal sensitivity in diabetic mice. Therefore, these results indicate that topical DNase I application promotes corneal epithelial wound healing and mechanical sensation restoration in diabetic mice, representing the potential therapeutic approach for diabetic keratopathy.

and expression disorders of regeneration-related signallings, such as phosphorylated epidermal growth factor receptor (EGFR), insulin-like growth factor 1 receptor (IGF-1R), silent information regulator 1 (Sirt1) and Akt. [6][7][8][9] Besides, increased inflammatory response is another considerable factor during the regeneration of corneal epithelium and nerve. Neutrophils are the first inflammatory cells migrating to the focal zone during inflammation or tissue damage. 10,11 One of the defence functions of neutrophils is the formation of neutrophils extracellular traps (NETs), a fibrous network consisting of nuclear and granular components. 12 The process of NETs formation and NETs release activation is defined as NETosis, 13,14 and previous studies have revealed that NETosis delays wound healing both in diabetic mice and humans. 15 It has been generally recognized that DNase families play vital functions in DNA degradation by targeting and shearing DNA sequences. There are two major DNase families, DNase I and DNase II, and multiple enzymes in each family play diverse roles in the development of various diseases. 16 Among these nucleases, DNase I has the widest description in DNase I family, 16 [21][22][23] and inflammatory disease. 24,25 Previous study have confirmed that DNase I-dependent NETs degradation was an important treatment for diabetic skin wound. 15 However, the NETs formation and the effects of DNase I in diabetic corneal epithelial and nerve regeneration remain unclear.
Commonly, the removal of extracellular DNA (eDNA) by DNase I administration and the application of NETs formation-related PAD4 (a calcium-dependent enzyme that is key in mediating NETosis) inhibitors (Cl-amidine) are two major methods for the intervention of NETs formation. 15,26 However, it still remains unclear about the function and mechanism of DNase I in diabetic corneal wound healing. Thereby, in this present study, we investigated the effects of DNase I and PAD4 inhibitors on corneal epithelial regeneration in diabetic mice and explored the functional mechanism of DNase I in diabetic corneal wound healing.

| Animals
Male C57BL/6 mice (6-8 weeks old) were purchased from the Institute of Laboratory Animal Sciences, Chinese Academy of Medical Sciences. All animal experiments were conducted with the approval of the Ethics Committee of Shandong Eye Institute according to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.
As described in our previous studies, type 1 diabetes was induced in the mice with intraperitoneal injections of streptozocin (STZ; 50 mg/ kg; Sigma-Aldrich). [27][28][29] Diabetic mice with a blood glucose level above 16.7 mmol/L at 16 weeks after the final STZ injection were used in this study. For each mouse, only one eye was suffered from the operation of epithelial removal.

| Corneal epithelial wound healing
Mice were anaesthetized by an intraperitoneal injection of 0.5 mL 0.9% sodium chloride solution containing 40 μL ketamine and 10 μL chlorpromazine hydrochloride, followed by topical application of proparacaine hydrochloride eye drops (Alcon) in the eye; then, an Algerbrush II remover (Alger Co) was used to scrape the central corneal epithelium (2.75 mm diameter) in mice. Fluorescein staining method was used to examine the residual epithelial defects, and the percentage of the original defect area was calculated with ImageJ software (NIH) as described in our previous studies. [29][30][31] For DNase I treatment, diabetic mice were topically applied with 1 mg/ mL DNase I (5 μL/eye; Roche Diagnostics GmbH) six times a day after the removal of corneal epithelium, whereas age-matched control mice were topically applied with equal phosphate buffer saline (PBS) as the vehicle control. For PAD4 inhibitor treatment, diabetic mice were intraperitoneal injected Cl-amidine (10 mg/kg; Absin Bioscience Inc) once a day from 5 days before corneal epithelial injury to 2 days after injury.

| Immunofluorescence staining
For immunofluorescence staining, mouse eyeballs were snap frozen with Tissue-Tek optimum cutting temperature compound (Sakura Finetechnical). Before staining, frozen sections (7 μm) were performed with pre-treatments including fixation (4% paraformaldehyde), permeabilization (0.1% Triton X-100) and blocking (10% bovine serum albumin), at room temperature. Then, processed samples were incubated with primary antibodies (showed in Table 1) overnight at 4°C and subsequently incubated with corresponding secondary antibodies for 1 hour at room temperature.
Information of primary and secondary antibodies could be addressed to Table 1. Besides, Hoechst 33342 (Solarbio) was used to stain DNA and Eclipse TE2000-U microscope was used for image capturing.

| Western blot
Total corneal protein was used for the detection of Western blot, which was extracted by lysis in radio immunoprecipitation assay (RIPA) buffer containing a proteinase inhibitor cocktail. Then, the extracted protein samples (10-20 µg total proteins) were run on 10% or 12% SDS-PAGE gels and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore). The protein-loaded membranes were incubated with primary antibodies (showed in Table 1) overnight at 4°C and then followed by the incubation of horseradish peroxidase (HRP)-conjugated secondary antibody (Amersham Biosciences) for 1 hour at room temperature. The target proteins were developed by the application of enzyme-linked chemiluminescence (ECL) kit (Pierce), and ImageJ software was used for the quantification of immunoreactive bands.

| Enzyme-linked immunosorbent assay
Mouse corneas were harvested at 48 hours after epithelial removal, and total corneal proteins were extracted with PBS. After centrifugation, sample supernatants were prepared for quantitative sandwich immunoassay detection with enzyme-linked immunosorbent assay (ELISA) kits, including myeloperoxidase (MPO; JEB) and neutrophil elastase (NE; JEB) according to the manufacturers' procedures.
Absorbance was read with a microplate reader (Molecular Devices) at 450 nm with a reference wavelength of 570 nm.

| Isolation and quantification of eDNA
Mouse corneas were harvested at 48 hours after epithelial removal, and eDNA was extracted with phenol (Sigma-Aldrich). 15 Then, the NanoDrop One spectrophotometer (Thermo Fisher Scientific) was used for measuring eDNA. Primer sequences for RT-qPCR were showed in Table 2.

| Measurement of corneal mechanical sensitivity
Corneal mechanical sensitivity was measured by using a Cochet-Bonnet esthesiometer (Luneau Ophtalmologie) without anaesthesia according to our previous studies. 28,31 The test began with the maximal length (6 cm) of nylon filament in the esthesiometer and shortened by 0.5 cm each time until we found the corneal touch threshold (which triggered the eye blink response). The longest filament length with a positive blink response was identified as the value of corneal sensitivity.

| Corneal whole-mount staining
Whole-mount corneas were fixed with Zamboni stationary liquid (Solarbio) for 1 hour; then, the cornea was dissected along the sclerallimbal region and blocked by PBS with 0.1% Triton X-100, 2% goat serum and 2% bovine serum albumin for 2 hour. Subsequently, Alexa Fluor 488-conjugated neuronal class III β-tubulin mouse monoclonal antibody was used for the incubation of whole-mount corneas overnight at 4°C. Eclipse TE2000-U microscope was used for the examination of nerve morphology and image capturing, and ImageJ

| Data analysis
All data in this study were obtained from at least three independent experiments, and results are presented as the means ± standard deviation (SD). Comparisons between two groups were performed using the Student t test, and a P value of less than .05 was used to indicate the statistical significance.

| RE SULTS
To assess the effects of DNase I on the regeneration of corneal epithelium, 1 mg/mL DNase I eye drops was administered to diabetic mice after the removal of corneal epithelium. Similar to previous studies, the regeneration rate of corneal epithelium delayed in diabetic mice, whereas DNase I application efficiently rescued the regeneration rate of corneal epithelium in diabetic mice ( Figure 1A).  . Data were given as the mean ± SD; *P < .05, **P < .01, ***P < .001 removal (24 hours: 23.6% ± 3.7% in healthy mice, 43.4% ± 10.5% in diabetic mice, 21.7% ± 4.7% in DNase I-treated diabetic mice; 48 hours: 0% in healthy mice, 10.9% ± 3.3% in diabetic mice, 0.9% ± 1.0% in DNase I-treated diabetic mice, Figure 1B; n = 5).
Besides, even there is no significant effect of Cl-amidine applica-  . C-E, Corneas harvested from healthy and diabetic (with or without 5-day DNase I treatment) mice were evaluated with Western blot to examine the protein levels of NADPH oxidase 2/4, and the quantified data of the Western blot results were shown (D, E; n = 6). Data were given as the mean ± SD; *P < .05, **P < .01, ***P < .001, n.s, not significant nerves ( Figure 6A,B; n = 5), both in the central and peripheral areas of diabetic corneas ( Figure 6B). Corneal mechanical sensitivity had significant improvement after 3 days' DNase I treatment in diabetic mice with epithelial removal, and there was no statistic differences between DNase I-treated diabetic mice and healthy controls at every test time-point ( Figure 6C; n = 5).

| D ISCUSS I ON
In this study, we found the eDNA-containing NETs were formed during corneal epithelial wound healing both in normal and diabetic mice. Elimination of eDNA by DNase I promoted the regeneration of corneal epithelium in diabetic mice, proving that eDNA was a negative factor in the process of corneal epithelium regeneration in diabetic mice. DNase I and PAD4 inhibitors are two major methods for the intervention of NETs formation, but which one is better is still unknown. Although PAD4 inhibitors promoted corneal epithelial regeneration in diabetic mice at 48 hours, they had no therapeutic effect at 24 hours. DNase I had an earlier onset of action than PAD4 inhibitors, which not only reduced the chance of infection after diabetic corneal injury, but also showed that eDNA was a more important factor affecting the healing of diabetic corneal injury than NETs formation. Therefore, we chosed DNase I for further detailed research. Oxidative stress is usually triggered by an imbalance between the formation of ROS and the antioxidant defence. 33 With hyperglycaemia as a major aetiological factor, the occurrence of diabetes has a high association with increased oxidative stress. 34 Previous studies have confirmed the effects of ROS accumulation in delaying wound F I G U R E 6 Effects of DNase I on the regeneration of diabetic corneal nerves and the restoration of mechanical sensation. A-C, Twentyone days after injury, the renewing corneas were harvested and the regenerated corneal nerve fibres were examined with corneal wholemount staining, and the immunofluorescence intensity of central (B) and peripheral (C) nerve fibres at 21 d after injury was calculated with ImageJ software (n = 5). D, A Cochet-Bonnet esthesiometer was used to test the mechanical sensitivity of the cornea in healthy, diabetic and DNase I-treated diabetic mice at 3, 7, 14 and 21 d after the corneal epithelial removal (n = 5). Data were given as the mean ± SD; *P < .05, **P < .01, ***P < .001, n.s, not significant healing of corneal epithelium, 8 and attenuation of ROS accumulation serves as an important therapeutic target for the treatment of delayed corneal regeneration in diabetic condition. 27,35 The formation of NETs depends on the ROS produced by NADPH oxidase, which means that oxidative stress is the key to neutrophil NETosis. [36][37][38] This present study consistently showed that DNase I application significantly alleviated the generation of ROS during diabetic corneal epithelial repair by suppressing the expression of NADPH oxidase. This was consistent with the results of Munafo DB's in vitro studies. 39 Inflammation is a necessary self-protective mechanism for organisms against the infection during unwarranted tissue damage, whereas the timely resolution of excessive inflammatory response is also required to maintain a healthy tissue-repair environment.

Results
A series of biological processes involve in the resolution of inflammation, which include but not limited to the suppression of neutrophil infiltration and the production of pro-inflammatory factors, promotion of macrophage phagocytosis and clearance of phagocytes. 40  can not only reduce the inflammatory response of neutrophils, but also that of macrophages.
Decreased corneal sensation and corneal nerve fibre density serve as the typical clinical characteristics of diabetic keratopathy. 45,46 Our and other previous studies have already manifested that reduced corneal innervation is an critical influencing factor for the impaired corneal wound healing in diabetic condition. 47 This work showed that topical DNase I administration promoted the regeneration of corneal nerve fibres and improved corneal sensation in diabetic mouse. There are few studies involving the relationship between eDNA and neurological diseases, except that NET markers are found in the brain parenchyma of Alzheimer's disease patients. 48 Therefore, our results suggest that DNase I administration is a potential therapy for the treatment of impaired corneal epithelial innervation, even diabetic neuropathy caused by hyperglycaemia.
In summary, the present study suggests that topical application of DNase I improves corneal epithelial wound healing, accompanied by the improvement of inflammatory resolution, corneal nerve fibre regeneration and corneal sensation restoration in diabetic mice.
The multiple function of DNase I in corneal epithelial and nerve regeneration provides a potential novel solution for the treatment of keratopathy in diabetic patients.

ACK N OWLED G EM ENTS
This work was partially funded by National Natural Foundation of . The funders had no role in study design, data collection, data analysis, interpretation, writing of the report or decision to submit the paper for publication.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest. All authors read and approved the final manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
All the data and materials generated and/or analysed during the current study are available.