Hyperosmolar Potassium Inhibits Corneal Myofibroblast Transformation and Prevent Corneal Scar

Abstract Purpose Corneal myofibroblasts play a crucial role in the process of corneal scarring. Potassium has been documented to reduce skin scar tissue formation. Herein, we investigated the ability of potassium to prevent corneal fibrosis in cell culture and in vivo. Methods Corneal fibroblasts (CFs) were isolated from the corneal limbus and treated with TGF-β1 to transform into corneal myofibroblasts. Corneal myofibroblast markers were detected by quantitative real-time PCR, Western blot, and immunofluorescence. The contractive functions of corneal myofibroblast were evaluated by the scratch assay and the collagen gel contraction assay. RNA sequencing in corneal fibroblasts was performed to explore the mechanisms underlying hyperosmolar potassium treatment. GO and KEGG analysis were performed to explore the underlying mechanism by hyperosmolar potassium treatment. The ATP detection assay assessed the level of cell metabolism. KCl eye drops four times per day were administered to mice models of corneal injury to evaluate the ability to prevent corneal scar formation. Corneal opacity area was evaluated by Image J software. Results Treatment with hyperosmolar potassium could suppress corneal myofibroblast transformation and collagen I synthesis induced by TGF-β1 in cell culture. Hyperosmolar potassium could inhibit wound healing and gel contraction in CFs. RNA sequencing results suggested that genes involved in the metabolic pathway were downregulated after KCl treatment. ATP levels were significantly decreased in the KCl group compared with the control group. Hyperosmolar potassium could prevent corneal myofibroblast transformation after corneal injury and corneal scar formation in mice. Conclusion Potassium can suppress corneal myofibroblast transformation and collagen I protein synthesis. Moreover, given that KCl eye drops can prevent corneal scar formation, it has been suggested to have huge prospects as a novel treatment approach during clinical practice.


Introduction
Corneal transparency mainly depends on its organizational structure, which involves a highly ordered arrangement of collagen fibers. 1,2 Normally, keratocytes are quiescent in the corneal stroma. However, they can differentiate into corneal fibroblasts (CFs) and myofibroblasts in response to different external stimulation, including trauma, infection, and inflammation. 3,4 Previous studies have demonstrated that corneal myofibroblasts play an essential role during the process of corneal fibrosis. 5,6 Corneal myofibroblasts can reportedly produce excessive amounts of collagen I, leading to the distortion of tissue architecture. 3,7 Consequently, inhibiting corneal myofibroblast transformation to prevent corneal fibrosis has become the focus of many studies.
To date, several approaches have been developed to inhibit corneal myofibroblast transformation. In our previous study, we demonstrated that inhibiting EZH2 could prevent corneal myofibroblast transformation. 8 It has also been shown that the p38 pathway may serve as a potential therapeutic target for human corneal fibrosis prevention. 9 Furthermore, raising cAMP by forskolin treatment reportedly inhibits transforming growth factor beta1 (TGF-b1)induced corneal myofibroblast transformation. 10 Last but not least, Rapamycin could inhibit the proliferation and differentiation of corneal myofibroblasts. 11 These results are indeed encouraging, however, there is currently a lack of preclinical and clinical studies to evaluate the effectiveness and safety of these drugs. Accordingly, the quest to find a new method to inhibit corneal transformation remains a high research priority.
Some studies have shown that potassium ion is an essential regulator of cell function. High extracellular potassium levels can reportedly preserve cell stemness by limiting nutrient uptake, 12 while potassium supplementation could affect the differentiation of human mesenchymal stem cells. 13 In addition, potassium gluconate treatment has been documented to reduce myofibroblasts within the dermis significantly probably directly through TGFb1 signaling. 14 While potassium has been shown to exert different effects in various cell types, its role in corneal fibroblast function and corneal tissue remodeling remains unclear.
In this study, we explored the antifibrotic effect of potassium on corneal fibroblast cells. Key genes involved in the process of potassium treatment were identified by RNA sequencing. We found that potassium could suppress the transformation of CFs into corneal myofibroblasts and inhibit corneal scar formation.

Ethics statement
Corneal tissues used in this study were acquired from Changsha Aier Eye Bank (Changsha, China). This study was approved by the Ethics Committee of the Aier Eye Hospital Group. Eight-week-old male and female C57BL/6 mice were provided by the Department of Laboratory Animals of Central South University. Animal procedures were approved by the Animal Welfare and Ethics Committee of Central South University. All mice were treated according to The Association for Research in Vision and Ophthalmology (ARVO) Statement.

Corneal fibroblasts isolation and culture
CFs were isolated and cultured as previously described. 8 Briefly, corneal tissue was digested for 60 min in Dispase II (Sigma-Aldrich, USA) at 37 C. Afterward, the epithelium and endothelium were carefully scraped using toothless forceps. Corneal tissue segments were decomposed into pieces, and then, digested overnight in collagenase (2 mg/mL) (Sigma-Aldrich, USA) at 37 C. CFs were collected and seeded into a six-well plate (Sigma-Aldrich, USA) and then cultured with DMEM/F-12 medium (Thermo Fisher Scientific, USA) supplied with 10% FBS (Thermo Fisher Scientific, USA). The cell culture medium was changed every other day. CFs were passaged when they reached 80-90% confluence.

Corneal fibroblasts cultured in KCl medium
The basic medium consisted of DMEM/F12 medium supplemented with 10% FBS. The KCl medium consisted of the basic medium with the addition of 60 mM KCl. CFs were seeded into 12-well plates and then treated with basic medium, TGF-b1 (2 ng/mL), KCl medium, and TGF-b1 (2 ng/mL) plus KCl medium for 72 h.

PCR
After treatment with different mediums for 72 h, CFs were harvested for quantitative real-time PCR (qPCR). Total RNA was extracted using Trizol reagent (Thermo Fisher Scientific, USA). After RNA extraction, the Reverse Transcription Kit (Vazyme Biotech, China) was used to reverse RNA transcription into cDNA. The DNA samples were then subjected to qPCR using the Vazyme qPCR kit (Vazyme Biotech, China). The gene primers used are described in Supporting Information Table 1.

Western blot
The protein expression of collagen I, fibronectin, a-SMA, SMAD3, PGD, P38, PTGES were tested by Western blot analysis. After CFs were cultured in the basic medium, TGF-b1 (2 ng/mL), KCl medium, and TGF-b1 (2 ng/mL) plus KCl medium for 72 h, cells were harvested for Western blot. Total protein was extracted using a protein extraction kit (Beyotime, China), according to the manufacturer's instructions. Proteins were separated by SDS gel electrophoresis. Then, proteins were transferred to the PDVF membrane (Milipore, USA). The proteins were revealed using chemiluminescence and photographed by an imaging system (LI-COR Biosciences, USA). Analyses of Western blot results were performed using ImageJ software (NIH Image J system, USA). Information on the antibodies used is described in Supporting Information Table 2.

Immunofluorescence staining assay
CFs were seeded in 24-well plates and then cultured using a basic medium, TGF-b1 (2 ng/mL), KCl medium, and TGF-b1 (2 ng/mL) plus KCl medium for 72 h. Subsequently, CFs were washed in sterile PBS three times and then routinely fixed in 4% paraformaldehyde for 20 min. Then, cell samples were permeabilized with PBS 0.5% Triton for 5 min. The cell samples were washed with sterile PBS three times and then blocked in 3% BSA in PBS (blocking solution). Afterward, the cells were incubated with the primary antibody at 4 C overnight and with the secondary antibody at room temperature for 1 h. Nuclei were counter stained with DAPI. A negative control was used to rule out nonspecific staining. Finally, photographs were taken using a Zeiss LSM microscope (Carl Zeiss, Germany).

Cell scratch assay
CFs were seeded in 12-well plates and then cultured in the basic medium, TGF-b1 (2 ng/mL), KCl medium, and TGF-b1 (2 ng/mL) plus KCl medium for 72 h. A 1 mL pipette tip was used to create a scratch wound when cells were grown to confluence. Afterward, the culture medium was replaced with DMEM/F12 medium containing 0.1% FBS. The wounds were photographed at 0 and 24 h after scratching. Quantification analysis of the gap area was performed using ImageJ software.

Gel contraction assay
CFs were seeded in 6-well plates and then cultured in the basic medium, TGF-b1 (2 ng/mL), KCl medium, and TGF-b1 (2 ng/mL) plus KCl medium for 72 h. CFs were digested with 0.25% trypsin and subsequently resuspended in DMEM/F12 medium. CFs were collected in a centrifuge tube and mixed with pre-cooling collagen I solution (Trauer Biotechnology, China). 15 The cells were seeded in 24-well plates at a density of 5 Â 10 5 /mL. The collagen gel was plated at 37 C for 30 min for solidification. Then, CFs were cultured in different medium as previously described. Contraction images of gels were photographed after treatment at 24 and 48 h. Quantification analysis of collagen gel sizes was performed using ImageJ software.

RNA-sequencing
CFs were treated with the basic medium (control group) and KCl medium for 72 h, each group consisted of four samples. RNA-sequencing was performed by SEQHEALTHtech Co., Ltd (Wuhan, China). The differentially expressed genes (DEGs) between the KCl group and control groups were identified using the edgeR filter criteria (j log2 (fold change) j > 1 and p-value < .05. GO analysis was performed using the GO database (http://www.geneontology. org/). KEGG pathway enrichment analysis was performed using the KEGG pathway database (http://www.genome.jp/ kegg). Visualization of enrichment results was performed using online OmicShare tools.

ATP detection assay
CFs were seeded into 6-well plates and then cultured in the basic and KCl medium for 72 h. Cells were harvested for ATP detection with an ATP detection assay Kit (Beyotime, China). Briefly, CFs were washed with sterile PBS twice. Then, lysis buffer was added to 6-well plates with 200 mL per well. The ATP detection reagent was diluted to 1:5 to obtain a working solution. 100 mL of the working solution was added to the 96-well plate for 5 min. The mixed solution was incubated for 5 min at room temperature. The results were analyzed with a multi-functional microplate reader.

Corneal injury and KCl eye drops treatment
The 15 mice were randomly divided into three groups. Mice were anesthetized with ketamine and a topical anesthetic applied to their ocular surface. Afterward, the corneal epithelium and anterior stroma were removed mechanically using the Algerbrush II (Ruijing technology, China). Group 1 mice were positive controls and did not receive any treatment. Group 2 mice were treated with 60 mM NaCl and group 3 mice were treated with 60 mM KCl. The uninjured left eye corneas were used as negative controls. Topical application of drugs was administered four times daily from postoperative day 4.
Slit-lamp biomicroscopy examination was performed on days 0, 4, 7, and 15 after injury to record the development of corneal opacity and epithelial defect. Quantification analysis of the corneal opacity area was performed using ImageJ software. Briefly, the scar area was manually described and then divided by the total corneal area. Then, the percentage of scar area was used to quantitative analysis.

Immunohistochemistry staining
Mice were sacrificed on day 15 post-injury. The eyeballs of mice were removed and fixed in 4% paraformaldehyde. Immunohistochemical staining was performed using a commercial IHC kit (Abcam, USA). Cryostat cornea sections were blocked in 3% hydrogen peroxide for 10 min at room temperature. After that, the protein block was incubated for 60 min. Afterward, the primary antibody was incubated overnight at 4 C and then incubated with secondary antibody for 60 min at room temperature. Streptavidin peroxidase was applied for 10 min at room temperature. Then, aminoethyl carbazole (AEC) single solution was applied for 10 min. Finally, tissue sections were incubated with hematoxylin for nuclear counterstaining.

Statistical analysis
Statistical analyses were performed using Welch's t-test and Welch's ANOVA tests. All data were graphed using GraphPad Prism (GraphPad Software, USA). A p value less than .05 was considered statistically significant.

Potassium suppressed corneal myofibroblast transformation induced by TGF-b1
We first sought to ascertain the optimal concentration of potassium. qPCR results showed that hyperosmolar potassium significantly suppressed COL1A1 expression induced by TGF-b1 in CFs (Figure 1(A)), while an equal concentration of NaCl had no effect (Figure 1(B)). However, CFs gradually died when the KCl concentration exceeded 60 mM (data not shown). Moreover, qPCR results revealed that hyperosmolar potassium could suppress the mRNA level of collagen I, fibronectin, and a-SMA (Figure 1(C)). The Western blot results of collagen 1, fibronectin, and a-SMA corresponded with qPCR results (Figure 1(D)). Meanwhile, immunofluorescence staining showed that hyperosmolar potassium could reduce fibronectin expression and a-SMA after TGF-b1 treatment (Figure 2). These results suggested that hyperosmolar potassium could inhibit corneal myofibroblast transformation in cell culture.

Potassium suppressed CF migration
To assess whether potassium could affect the migration ability of CFs, the cell scratch and gel contraction assays were performed. The cell scratch assay results revealed that hyperosmolar potassium could suppress wound healing in CFs (Figure 3(A,B)). The collagen gel contraction assay indicated that hyperosmolar potassium could inhibit collagen gel contraction (Figure 3(C,D)). These results suggested that potassium could suppress fibroblast transformation induced by TGF-b1.

Potassium suppressed SMAD and p38 MAPK pathway
According to previous reports, the SMAD and p38 MAPK signaling pathways are involved in the process of corneal fibrosis. 9 To determine the effect of potassium on these pathways, we assessed the protein levels of SMAD3 and p38. The Western blot results revealed that potassium significantly inhibit the protein level of SMAD3 and p38 (Supporting Information Figure S1A). These results suggested that potassium could inhibit corneal fibrosis by suppressing SMAD and p38 MAPK pathway.

RNA-sequencing analysis results of potassium treatment
To further study the molecular mechanism underlying the effects of potassium, RNA-seq analysis was performed. Principal component analysis (PCA) results revealed significant differences between the KCl and control groups ( Figure  4(A)). Pearson correlation analysis results demonstrated good repeatability of KCl treatment (Figure 4(B)). The differential gene expression results were visualized using a volcano plot. The upregulated genes in KCl group included FGF1, MAMDC2, DKK1, CRYAB, and SLC5A3, downregulated gene included NDNF, HSPB6, GAP43, PTGES, and TXNIP (Figure 4(C)). The screening identified 721 upregulated DEGs and 795 downregulated DEGs between the KCl and control groups. A heatmap was generated to display differential gene expression in the two groups (Figure 4(D)).

GO enrichment analysis and KEGG pathway analysis
The top 20 significantly enriched GO terms of the DEGs were selected (Figure 5(A)). The GO terms are provided in Supporting Information Table 3. The significantly enriched GO terms were related to fibrosis progression, including response to metabolic and cellular processes and regulation of biological processes.
We further analyzed the enrichment of KEGG pathways based on DEGs. Figure 5(B) shows the top 20 significantly enriched pathways of upregulated DEGs, including regulation of actin cytoskeleton, Rap1 signaling pathway, PI3K-Akt signaling pathway, MAPK signaling pathway, Gap junction, Cell cycle, and apoptosis. Figure 5(C) shows the top 20 significantly enriched pathways of downregulated DEGs, including TNF signaling pathway, TGF À beta signaling pathway, PI3K À Akt signaling pathway, Metabolic pathways, Focal adhesion, ECM À receptor interaction, and Cell adhesion molecules (CAMs). Metabolic pathways were the most significantly enriched.

Potassium suppressed CF fibrosis by inhibiting cell metabolism
KEGG pathway analysis yielded three significantly enriched pathways associated with fibrosis metabolic pathways, Rap1 signaling pathway, and apoptosis. Figure 6(A) illustrates the expression levels of DEGs in these pathways. Five representative genes from these pathways were selected to validate the gene expression according to RNA-sequencing results (Figure 6(B)). qPCR analysis results were in accordance with the RNA-sequencing results. Gene expressions of PGD and PTGES in the KCl group were significantly lower than in the control group; FGF1, HGF, and ATF4 were higher than in the control group. The Western blot results of PGD and a-PTGES corresponded with qPCR results (Supporting Information Figure S1A). Meanwhile, the intracellular ATP levels in the KCl group were significantly lower than in the control group (Supporting Information Figure S1B). Overall, our findings suggest that potassium could inhibit cell metabolism in CFs.

PGD inhibitor suppressed the expression of myofibroblast markers induced by TGF-b1
To determine the optimal dose of physcion, 16 we tested different concentration of physcion on CFs. qPCR results revealed that 5 lM physcion was the optimal concentration for the suppression of PGD expression (Supporting Information Figure S1C). Physcion (5 lM) suppressed the expression of collagen 1, fibronectin, a-SMA induced by TGF-b1 (Supporting Information Figure S1D). These results suggested that PGD inhibitor could inhibit corneal myofibroblast transformation in cell culture.

Potassium prevented corneal scar formation in injury mice model
To study the efficacy of potassium in preventing corneal scar formation, KCl and NaCl eye drops were administered topically to mice models of corneal injury. Corneal brightfield photographs were captured at days 0, 4, 7, and 15 postinjury (Figure 7(A)). Our results indicated that corneal opacity in the KCl group was significantly lower than in the untreated injured group, while the NaCl group had no effect (Figure 7(B)). Immunostaining results showed that a-SMA was significantly upregulated in injured mice cornea ( Figure  8(A)). The expression of a-SMA in the KCl group was lower than in the untreated group, while no significant changes were observed in the NaCl group (Figure 8(B)). Overall, our results revealed that hyperosmolar potassium could prevent corneal myofibroblast transformation after corneal injury and corneal scar formation.

Discussion
Corneal fibrosis comprises a complex series of related events, which involve cell migration of epithelial, stromal, and endothelial cells, regulation by growth factors, and collagen remodeling. 5 To date, the molecular mechanisms of corneal fibrosis remain largely unknown. Metabolic alterations are increasingly recognized as critical pathogenic processes that underlie fibrosis across many organ types. [17][18][19][20] Moreover, inhibition of glycolysis by the PFKFB3 inhibitor blunted the differentiation of lung fibroblasts into myofibroblasts and attenuated profibrotic phenotypes in pulmonary myofibroblasts. 19 In addition, inhibition of Acetyl-CoA carboxylase and de novo lipogenesis was reported to directly suppress the activation of hepatic stellate cells and prevent liver fibrosis. 21 Furthermore, it has been shown that myokine-mediated muscle-kidney crosstalk can suppress metabolic reprogramming and fibrogenesis in kidney diseases. 22 As a result, metabolically targeted therapies could become essential strategies for the reduction of fibrosis.
A previous study that demonstrated the effectiveness of potassium gluconate treatment on skin wound healing suggested that it could be an efficient treatment to prevent scar formation. 14 Consistently, we found that KCl could inhibit corneal myofibroblast transformation, and the antifibrotic effect of KCl was dose-dependent. However, no changes were observed in the NaCl treatment group, suggesting that potassium is an effective critical component of the antifibrotic effect. Besides, our results showed that KCl eye drops could prevent corneal scarring. KCl is a metal ion that can easily penetrate the corneal epithelial barrier. Therefore, the application of KCl eye drops was selected instead of subconjunctival injections. In addition, the effect of medication of our eye drops in other days is worth further study.
To elucidate the antifibrotic mechanism of potassium, we conducted an RNA sequencing study. RNA sequencing results revealed that 1516 genes were differentially expressed between the KCl and control groups. The number of upregulated and downregulated DEGs was similar. According to the results of the top 20 significantly enriched KEGG pathways, we focused on metabolic pathways, Rap1 signaling pathway, and apoptosis pathway. PGD and PTGES were selected from the metabolic pathways. PGD (6-Phosphogluconate dehydrogenase) is an key enzyme in the pentose phosphate pathway, then, indirectly involved with ATP productin in glycolysis pathway. 23 An increasing body of evidence suggests that PGD is upregulated in many solid cancers, [24][25][26] and PGD activity inhibition can inhibit cell proliferation and tumor growth. 27 In addition, in a rat model of pulmonary fibrosis induced by bleomycin, upregulated expression of PGD was found, and PGD activity inhibition resulted in an antifibrotic effect. 28 Furthermore, studies have shown that knocking out PTGES can promote liver repair after liver injury in mice. 29 PTGES (Prostaglandin E Synthase) is an enzyme in the prostaglandin E synthase pathway. ATP could stimulate arachidonic acid release and the synthesis of prostaglandin E2. 30 In the present study, PGD and PTGES in the KCl group were significantly downregulated compared with the control group, suggesting that potassium may exert an antifibrotic effect by inhibiting the expression of PGD and PTGES via regulating ATP production. In addition, our results showed that PGD inhibitor could inhibit corneal myofibroblast transformation in cell culture. These results suggested that metabolic intervention could regulate the process of corneal fibrosis. Besides, our results showed that ATP levels were significantly decreased in the KCl group, indicating that potassium may inhibit corneal myofibroblast transformation by inhibiting cell metabolism.
Furthermore, FGF1 and HGF were selected from the Rap1 signaling pathway. Interestingly, studies have shown that FGF1 can inhibit the epithelial-mesenchymal transition (EMT) of glomerular podocytes. 31 Moreover, FGF1 effectively inhibits the EMT of lung epithelial cells induced by TGFb1 through the MAPK/ERK kinase pathway. 32 Furthermore, HGF can reportedly prevent or reverse fibrosis in heart injury models. 33 In this study, FGF1 and HGF were upregulated in the KCl group, and ATF4 was identified from the apoptosis pathway. ATF4 has been reported to prevent liver steatosis in mice by inducing SIRT2 expression. 34 Our results also showed that ATF4 was upregulated in the KCl group, suggesting that potassium may upregulate these antifibrotic genes to inhibit myofibroblast transformation.
Besides, several studies have found a close relationship between the potassium channels and fibrotic diseases. Nattel et al. showed that fibroblast KCNJ2 expression and currents are upregulated in congestive heart failure, thereby hyperpolarizing resting membrane potential and enhancing atrial fibroblast proliferation. 35 Bradding et al. demonstrated that the potassium channel K(Ca)3.1 plays a crucial role in human fibrocyte migration. 36 Moreover, K(Ca)3.1 inhibitor has been documented to significantly attenuate corneal fibrosis in cell culture. 37 Accordingly, we hypothesize that hyperosmolar potassium could affect the function of the potassium channel in CFs, and hence, inhibit cell fibrosis. Further studies are needed to explore the mechanisms underlying the antifibrotic function of potassium.
The present study had a few limitations. During the animal experiments, the potassium levels in the anterior chamber or blood were not quantified. Given that it has been established that external potassium can cause arrhythmias, 38 the safety of KCl eye drops warrants further study. Besides, our results only showed that potassium treatment could prevent scar formation in mice models of corneal injury. However, little is known about the ability of KCl eye drops to treat existing corneal scars. In this study, we demonstrated that hyperosmolar potassium could inhibit SMAD and p38 expression, which were key role in TGF-b1 pathway. 39,40 The underlying mechanism of hyperosmolar potassium treatment regulate SMAD and p38 need for further investigation.
In summary, potassium could inhibit corneal myofibroblast transformation induced by TGF-b1. The antifibrotic mechanism of potassium may involve the inhibition of cell metabolism. Moreover, we demonstrated that potassium could prevent corneal scar formation in mice models of corneal injury. Overall, our results indicated that potassium could serve as a novel approach for preventing corneal scar formation.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This study was supported by the grant from National Natural Science Foundation of China [81871495]. Authors have no commercial interest to disclose.