Nitric oxide attenuated transforming growth factor-β induced myofibroblast differentiation of human keratocytes

Nitric oxide (NO) has the potential to modulate myofibroblast differentiation. In this study, we investigated the effect of exogenous NO on the myofibroblast differentiation of human keratocytes using sodium nitrite as a NO donor. Myofibroblasts were induced by exposing resting keratocytes to transforming growth factor (TGF)-β1. N-cadherin and α-smooth muscle actin (αSMA) were used as myofibroblast markers. Both resting keratocytes and -stimulated keratocytes were exposed to various concentrations of sodium nitrite (1 μM to 1000 mM) for 24 to 72 h. Exposure to sodium nitrite did not alter keratocytes’ viability up to a 10 mM concentration for 72 h. However, significant cytotoxicity was observed in higher concentrations of sodium nitrite (over 100 mM). The expression of αSMA and N-cadherin was significantly increased in keratocytes by TGF-β1 stimulation after 72 h incubation. The addition of sodium nitrite (1 mM) to TGF-β1-stimulated keratocytes significantly decreased αSMA and N cadherin expression. Smad3 phosphorylation decreased after sodium nitrite (1 mM) exposure in TGF-β1-stimulated keratocytes. The effect of NO was reversed when NO scavenger, 2-4-carboxyphenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (cPTIO) was added in the culture medium. Application of sodium nitrite resulted in significant decrease of corneal opacity when measured at 2 weeks after the chemical burn in the mouse. These results verified the potential therapeutic effect of NO to decrease myofibroblast differentiation of human keratocytes and corneal opacity after injury.

The results of previous studies suggest the potential therapeutic effect of NO in the healing process of corneal wounds. Recently, the application of exogenous NO in the ophthalmic field was actively investigated. The permissive role of NO in corneal epithelial wound healing was reported previously [13][14][15][16][17] . The topical application of NO successfully promoted the corneal epithelial wound healing process 13,17 . In addition, NO's antibacterial effect is another benefit to prevent further corneal damage from bacterial infection after injury 18,19 . Incidentally, a recent development of NO as a promising anti-glaucoma medication further increased the clinical interest of NO in the ophthalmic field 9,20,21 . Although NO ameliorates corneal epithelial wound healing, corneal injury usually results in both corneal epithelial and stromal damage simutaneously. Of course, keratocytes are the major cell component of corneal stroma. In corneal scars, keratocytes differentiate into myofibroblasts and lay down abnormal collagen fibers that can deteriorate corneal transparency. Therefore, the modulation of myofibroblast differentiation in an injured cornea is a critical therapeutic target to minimize corneal opacity and preserve clear vision. From this perspective, the evaluation of the effect of NO on keratocytes is a necessary step for the development of NO as a corneal wound healing modulator. Although the anti-fibrotic action of exogenous NO was reported in various human tissues, its role in corneal fibrosis, especially myofibroblasts' differentiation from keratocytes, has not been fully elucidated. If NO is found to benefit both corneal epithelial cells and stromal cells, the further development of new drugs using NO can be more effective.
In this study, we investigated the effect of exogenous NO on primarily cultured human keratocytes. Different concentrations of NO donors (sodium nitrite) were applied in the culture media, and the cellular viability of keratocytes was measured. We induced keratocytes' myofibroblast differentiation by adding transforming growth factor β1 (TGF-β1) to the culture media and investigated the effect of NO on myofibroblast markers' expressions, N cadherin and α-smooth muscle actin (αSMA) from TGF-β1-stimulated keratocytes. Finally, the effect of NO on corneal opacity development was evaluated using murine chemical corneal burn model.

Results
Keratocytes' viability with different concentrations of NO donors. We investigated any toxic effect of NO on keratocytes. Keratocyte viability was not deteriorated at low concentrations (up to 10 mM) of sodium nitrite. Rather mild increase of cell viability was observed with the addition of 10 mM of sodium nitrite. However, sodium nitrite decreased keratocytes' viability at high concentrations (equal to or more than 100 mM). This toxicity increased with a longer incubation period. The decrease of viability in high concentrations (over 100 mM) of sodium nitrite is attributed to the hyper-osmolar stress induced by excess sodium in the culture media ( Fig. 1). Intracellular NO concentration after exposure to different concentrations of sodium nitrite was measured (Fig. 2). A mild increase of intracellular NO concentration was observed after 24, 48 and 72 h incubation. Addition of 2-4-carboxyphenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (cPTIO, 10 μM) in the culture medium scavenged NO and decreased intracellular NO concentration. NO activates guanylate cyclase and increase the production of cGMP. Significant increase of intracellular cGMP was observed after exposure to sodium nitrite. No significant change of oncogene activation related proteins, p53 and p21 was observed with the NO donor treatment keratocytes. (Supplement Fig. S1). mTOR pathway activation and cellular autophagy. Two critical cell survival pathways (mTOR and autophagy) were not affected by sodium nitrite up to a 1000 μM concentration after 24-h exposure (Fig. 3).
Nitrite reductase activity in keratocyte. Nitrite reductase activity in keratocyte was measured in keratocytes using mitochondrial amidoxime reducing component 1 (mARC1), xanthine dehydrogenase and xanthine oxidase. The mRNA expression level of xanthine dehydrogenase was significantly increased after TGFβ1 stimulation while the mRNA expression level of mARC1 showed little change. The protein level of xanthine oxidase was also increased after TGFβ1 stimulation (Fig. 6).
NO's effect on αSMA expression from TGFβ1-stimulated keratocytes. We evaluated the effect of sodium nitrite treatment on TGFβ1-stimulated keratocytes. We added various concentrations (100 and 1000 μM) of sodium nitrite into the culture media. The addition of 1000 μM of sodium nitrite significantly decreased Smad3 phosphorylation, N-cadherin and αSMA expression in TGFβ1-stimulated keratocytes. However, lower concentrations (100 μM) of sodium nitrite failed to show any significant effect (Fig. 7). Addition cPTIO (10 μM) in the culture medium scavenged NO effect. The effect of sodium nitrite on αSMA expression in TGFβ1-stimulated keratocytes was further verified by using a different NO donor, DETA NONOate (Supplement Fig. S2). Inhibition of soluble guanylate cyclase by ODQ successfully reversed αSMA expression in TGFβ1stimulated keratocytes (Fig. 8). This finding suggests that the role of cGMP is critical in NO mediated prevention of myofibroblast differentiation from keratocytes.
In vivo effect of NO on corneal opacity development after injury. Topical treatment with sodium nitrite (100 μM and 1000 μM) significantly decreased the corneal opacity development after alkali burn. The www.nature.com/scientificreports/ effect of both concentrations of sodium nitrite was similar. Histologic examination of revealed decreased corneal edema and cellularity (representing residual corneal inflammation) in corneal stroma in NO treated eyes compared to PBS control ( Fig. 9).

Discussion
In this study, we investigated the effect of NO on myofibroblast differentiation from human keratocytes. The safety of NO with keratocytes was verified by the maintenance of cell viability and intact mTOR/autophagy pathways observed up to 1 mM of sodium nitrite. The addition of 1 mM of sodium nitrite in the culture media resulted in significant decrease of the intracellular cGMP level, expression of αSMA and N-cadherin from TGFβ1-stimulated human keratocytes. Blocking the cGMP formation by soluble guanylate cyclase inhibitor (ODQ) reversed the NO effect on αSMA expression. Application of topical NO in the setting of chemical corneal burn resulted in significantly decreased corneal opacity. Transforming growth factor-β1 (TGF-β1) is a strong inducer of myofibroblast differentiation through a prooxidant shift in redox homeostasis, which is associated with decreased NO/cGMP signaling 10 . ROS derived from NADPH oxidase (NOX4) mediated αSMA and collagen production by intestinal or nasal myofibroblasts when stimulated with TGF β 22,23 . After corneal injury, it is known that TGF-β1 produced in corneal epithelial cells can leak through the break of the epithelial basement membrane and activate keratocytes into myofibroblasts 24 . Therefore modifying NO/cGMP signaling pathway can be a potential therapeutic target to minimize corneal opacity in pathologic condition. It is reported that decreases of cGMP level at the wounding site can drive the myofibroblast differentiation of dermal fibroblasts. Conversely, the combinatorial effect of activators of soluble guanylate cyclase and inhibitors of cGMP degradation may lead to an elevation of cGMP signaling and induce the reversal of myofibroblast differentiation, as demonstrated in prostatic and dermal stromal cells 10 . As mentioned earlier, NO activates soluble guanylate cyclase and increases cGMP level in cells 9 . Therefore, the result of our current study is in line with previous reports about inhibitory NO effect on myofibroblast differentiation. In addition, as shown in Fig. 5, the increase of ROS can be alleviated by NO treatment.
Myofibroblasts play important roles in the corneal wound healing process [24][25][26] . Immediately after corneal injury, various cytokines, growth factors, and chemokines orchestrate the corneal wound healing process 26 . During the acute phase of corneal injury, damaged corneal epithelial cells produce profuse pro-inflammatory cytokines, such as interleukin 1 (IL-1), TGF-β1, and platelet-derived growth factor (PDGF), and these cytokines induce apoptosis of keratocytes at the injured area 27 . After the acute phase, keratocytes from the adjacent corneal stroma start to proliferate and migrate into the injured area. TGF-β1 can generate myofibroblasts from activated keratocytes 24 . The keratocyte derived myofibroblasts express several intermediate filament proteins, such as αSMA, vimentin, and desmin, which are important for providing mechanical strength to the injured tissue 28 . In addition, disorganized collagen fibers are produced by activated myofibroblasts. These abnormal collagen fibers and their irregular arrangement are the major causes of corneal opacity after wound healing.
Our finding that NO could decrease αSMA expression from activated keratocytes has an important clinical relevance, because the increased expression of αSMA is considered the key step toward myofibroblast www.nature.com/scientificreports/ differentiation. We tested various concentrations of NO donor, sodium nitrite, in the current study, because NO's biologic effect is known to be dependent on its concentration. Previously, lower concentrations of NO were reported to exert a direct positive effect on various cellular proliferations, whereas higher concentrations of NO may have cytotoxic effects possibly through both oxidation and nitrosative stresses 3,29 .
We found that exogenous NO could prevent myofibroblasts differentiation with little harmful effect on keratocyte's viability. This finding is an important clue that NO releasing treatment platforms can be safely used in various corneal traumatic or infectious diseases. We previously reported that exogenous NO can facilitate cornea  Total reactive oxygen species (ROS) production was measured in keratocytes after 24, 48 and 72 h culture with TGFβ1 (10 ng/mL). Slight increase of total ROS was observed in TGFβ1 (10 ng/mL) stimulated keratocytes compared to control keratocytes (Ctrl). Treatment with sodium nitrite (10, 100 or 1000 μM) alleviated total ROS production in stimulated keratocytes. **p < 0.01, ***p < 0.01.  30 . Haze control after photorefractive keratectomy can be another implication of exogenous NO treatment because activation of stromal myofibroblasts was known as the main cause of delayed corneal haze after PRK 25 . Although mitomycin C (MMC) is wildly used to prevent corneal haze in the clinical setting, the long term safety issue of MMC has yet to be verified 31,32 . An ideal therapeutic agent to prevent post PRK corneal haze should be safe for keratocytes and can specifically block myofibroblast differentiation. In this respect, NO can be a promising candidate when considering its safety and efficacy. Sodium nitrite not only stimulates cGMP production, but also promotes S-nitrosylation of cellular proteins by forming RSNO 33,34 . As shown in our study (Supplement Fig. S3), addition of sodium nitrite in TGFβ1-stimulated keratocytes significantly increased RSNO production, which can increase S-nitrosylation of corneal proteins. Investigation of protein functional change by NO induced S-nitrosylation in cornea can be an interesting topic for the future study.  www.nature.com/scientificreports/ Our study has some limitations. The first is the relatively short-term effect of NO on keratocytes that we observed. Because corneal opacity usually develops 2 to 4 weeks after injury in human, our in vitro data cannot provide the long-term effect of NO on the myofibroblast differentiation of keratocytes and its maintenance. Although chemical burn model was produced in a mouse model, it is possible that the wound healing response of mouse can be different from human. Therefore, the result is not always repeatable in human because of the many confounding factors of the human ocular surface. Another limitation is that sodium nitrite was used as an NO donor. Sodium nitrite is one of the widely accepted NO donors in various experimental settings. However, Figure 8. Evaluation of cGMP dependent pathway using guanylate cyclase inhibitor, ODQ (20 μM), in TGFβ1 (10 ng/mL) stimulated keratocyted. Addition of ODQ eliminated the anti-myofibroblastic effect of NO donors (DETA NONOates and sodium nitrite) as demonstrated by the restored α-SMA expression. Statistical significance was determined using one-way ANOVA followed by the Bonferroni multiple comparison test and the significant difference compared to no treated control (*p < 0.05, **p < 0.01); # significant difference compared to TGFβ1 treated group ( # p < 0.05). www.nature.com/scientificreports/ NO's release from sodium nitrite necessitates nitrite reductase. Therefore, the final NO concentration after sodium nitrite exposure largely depends on cellular nitrite reductase levels. Although we verified sodium nitrite reductase activity in resting keratocytes and its increase with TGFβ1 stimulation, accurate titration of NO supply was impossible with the current setting of experiment. For a more accurate titration of the NO supply, it is more desirable to develop enzyme-independent NO delivering platforms in the future.

Scientific Reports
In conclusion, we found exogenous NO prevented myofibroblastic differentiation from TGF-β1-stimulated human keratocytes. These findings suggest the future use of exogenous NO-releasing drug platforms for the treatment of various ocular diseases threatening corneal transparency.

Materials and methods
Human keratocytes' culture. The primary culture of human keratocytes was performed performed using a cadaveric donor corneal tissue not suitable for clinical use (from Eversight Korea, Seoul, South Korea) as described earlier 35 . Informed consent was obtained from next of kin of the cadaver donor for the tissue to be used in research. The use of human keratocytes was approved by the Institutional Review Board of Dongguk University Ilsan Hospital (IRB No: 2019-03-001) and the research was performed in accordance with the Declaration of Helsinki. Briefly, Descemet's membrane and epithelium were gently removed using forceps and an ophthalmic knife from the donor corneal button. The corneal stroma was minced in a laminar flow hood. Subsequently, mid-stroma and posterior stroma explants were suspended in a culture medium and cultured in 24-well plates. The corneal stroma was sliced into quarters and digested overnight with 2.0 mg/mL collagenase (Roche Applied Science, Mannheim, Germany) and 0.5 mg/mL hyaluronidase (Worthington Biochemicals, Lakewood, NJ, USA) in DMEM at 37 °C. Isolated cells were washed in DMEM and cultured in Keratinocyte SFM (Gibco BRL, Carlsbad, CA, USA). The cells were cultured on tissue culture-treated plastic at 4 × 10 4 cells/cm 2 . After reaching confluency, cells were harvested and suspended in a culture medium. The cells were plated in 75 cm 2 tissue flasks and maintained at 37 °C in 5% CO 2 and 95% air. The culture medium was changed every three days, and the cells were passaged using 0.25% Trypsin-EDTA (Gibco BRL, Carlsbad, CA, USA). Passage numbers 5-7 were used in this study.
Cell viability assay. Cell viability assays were performed using a cell counting kit reagent (CCK-8; Dojindo Molecular Technologies, Inc., Kumamoto, Japan) according to the manufacturer's protocol 35 . Briefly, keratocytes were cultured at 3 × 10 3 cells per well in a 96-well plate and incubated for 24 h. Following the adherence of cells, different concentrations of sodium nitrite were added to the culture media for 24 h, 48 h, and 72 h. The wells with no sodium nitrite and the wells with a dimethyl sulfoxide (DMSO) addition were used as negative and positive controls, respectively. After the appropriate incubation, 10 uL of CCK-8 solution was added to each cultured well, and the absorbance was measured at 450 nm after 2 h incubation of keratocytes with the reagent. Cell viability in various sodium nitrite solutions was represented as a relative percentage compared to the negative control. Reactive oxygen species (ROS) assay. Total ROS was detected using DCFDA/H2DCFDA-Cellular ROS Assay Kit (Cat. No. ab113851: Abcam, Cambridge, UK). Following treatment of TGFβ1 (10 ng/mL) or NaNO 2 (10, 100, 1000 μM), cellular ROS was measured at 24 h, 48, and 72 h time point. As mentioned in manufacturer's protocol, the cells were stained with 20 µM of 2′,7′-dichlorofluorescin diacetate (DCFDA) solution and incubated at 37 °C in the dark condition for 45 min. After discarding the solution, 100 µL/well of 1× buffer was added. The fluorescence was immediately measured at 485 nm excitation/535 nm emission.
Measurement of total s-nitrosothiol. The measurement of total S-nitrosothiols (RSNO) was carried out using Griess/Saville method. Briefly, TGFβ1 (10 ng/mL) stimulated keratocytes were cultured for 24 h with or without sodium nitrite (0, 100, 1000 μM). The cell lysates were mixed with equal volumes of 1× Griess reagent (catalog number: G4410, Sigma Aldrich) which was freshly prepared with or without 3 mM HgCl 2 . The absorbance was measured at 540 nm and the amount of RSNO was calculated using a standard curve. www.nature.com/scientificreports/ Statistical analysis. Data are presented as mean ± standard error, and statistical significance was determined using one-way analyses of variance (ANOVA) and Dunnett's multiple comparison tests at cGMP assay or western blot analysis and two-way ANOVA followed by the Bonferroni multiple comparison test also carried out at the NO detection analysis. In this study, p < 0.05 was regarded as significant, and calculations were completed with GraphPad Prism v5.01 (GraphPad Software, Inc., La Jolla, CA, USA).

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.