Acetylcholine decreases formation of myofibroblasts and excessive extracellular matrix production in an in vitro human corneal fibrosis model.

Abstract Acetylcholine (ACh) has been reported to play various physiological roles, including wound healing in the cornea. Here, we study the role of ACh in the transition of corneal fibroblasts into myofibroblasts, and in consequence its role in the onset of fibrosis, in an in vitro human corneal fibrosis model. Primary human keratocytes were obtained from healthy corneas. Vitamin C (VitC) and transforming growth factor‐β1 (TGF‐β1) were used to induce fibrosis in corneal fibroblasts. qRT‐PCR and ELISA analyses showed that gene expression and production of collagen I, collagen III, collagen V, lumican, fibronectin (FN) and alpha‐smooth muscle actin (α‐SMA) were reduced by ACh in quiescent keratocytes. ACh treatment furthermore decreased gene expression and production of collagen I, collagen III, collagen V, lumican, FN and α‐SMA during the transition of corneal fibroblasts into myofibroblasts, after induction of fibrotic process. ACh inhibited corneal fibroblasts from developing contractile activity during the process of fibrosis, as assessed with collagen gel contraction assay. Moreover, the effect of ACh was dependent on activation of muscarinic ACh receptors. These results show that ACh has an anti‐fibrotic effect in an in vitro human corneal fibrosis model, as it negatively affects the transition of corneal fibroblasts into myofibroblasts. Therefore, ACh might play a role in the onset of fibrosis in the corneal stroma.


| 4851
SŁONIECKA ANd dANIELSON conjunctiva and lacrimal gland. Upon injury to the cornea that extends deeper than the epithelium, a layer of randomly arranged collagen fibres (called Bowman's layer), which normally limits the passage of TGF-β1 further down to the stroma, is destroyed. This event results in TGF-β1 penetration into the stroma and initiation of the wound healing process. 7 Transforming growth factor-β1 induces excessive production of ECM components such as collagens I and III [8][9][10] and fibronectin (FN) 11,12 by activated fibroblasts.
Moreover, TGF-β1 promotes differentiation of fibroblasts into myofibroblasts. 13 In contrast to keratocytes, which are quiescent cells of the corneal stroma, and which main function is to sustain components of the ECM, 14 myofibroblasts produce strong contractile force in order to close the injured tissue, 15 and they express markers such as alpha-smooth muscle actin (α-SMA), vimentin and desmin. 16 Additionally, myofibroblasts are opaque due to decreased expression of corneal crystallins, such as aldehyde dehydrogenase class 1 (ALDH1), 17 and produce disorganized ECM, which makes them responsible for corneal haze [17][18][19] and lowered mechanical properties of the cornea. 20 Contractile activity and expression of α-SMA decrease when the injured tissue is properly healed, and myofibroblasts undergo apoptosis. 21 However, as mentioned above, in pathological process of healing, the myofibroblasts do not cease their activity, which in turn leads to formation of fibrotic tissue. 16 Acetylcholine (ACh) has been regarded as a classical neurotransmitter, released by cholinergic neurons and acting through activation of nicotinic and muscarinic receptors (n-and mAChRs). 22 However, for the past decades, increasing evidence has shown that ACh is synthesized by a majority of human cells and that it modulates various cellular processes. [23][24][25][26] For example, stimulation of ACh receptors has been found to have an anti-inflammatory effect, [27][28][29] to induce proliferation markers in keratinocytes 30 and to stimulate skin wound healing. 30 In the cornea, the corneal epithelium has one of the highest concentrations of ACh in the body, 31,32 and it has been suggested that ACh might accelerate corneal re-epithelialization 33,34 and play a role in migration of corneal epithelial cells. 35 On the contrary, the concentration of ACh in the corneal stroma is very low, but our previous studies have shown that the resident keratocytes are able to produce and secrete ACh in in vitro settings. 36 We have also demonstrated that ACh induces proliferation of keratocytes 37 and that it decreases keratocyte apoptosis in a Fas-ligand apoptosis model. 38 Research on the role of ACh in fibrosis has been mostly conducted on its role in airway diseases, for which it has been shown that stimulation of mAChRs may be involved in remodelling processes in chronic airway diseases 39 and that α7 nAChR is a key regulator of lung fibrogenesis. 40 It has also been shown that ACh induced collagen expression and proliferation of myofibroblasts in hepatic stellate cells. 41 At present, however, it is not known what effects, if any, ACh has on the differentiation of corneal fibroblast into myofibroblasts, nor on the development of fibrosis in the cornea. In this project, we studied the role of ACh in the transition of corneal fibroblasts into myofibroblasts, and therefore its role in the onset of fibrosis, in an in vitro human corneal fibrosis model.

| Human corneas
Healthy human corneas, which were obtained from deceased individuals who had chosen, when alive, to donate their corneas for transplantation or research purposes according to the Swedish law, were stored in the Tissue Establishment, Eye Bank Umeå, at the University Hospital of Umeå, Sweden, and delivered to the research laboratory if they were not used for transplantation. The project was vetted by the Regional Ethical Review Board in Umeå, which determined it to be exempt from the requirement for approval (2010-373-31M). The study adhered to the tenets of the Declaration of Helsinki.

| Isolation and culture of primary keratocytes
Primary keratocytes were isolated from 14 donors. The isolation and culture of primary keratocytes have been described previously. 36 Shortly, in order to remove remaining epithelial and endothelial cells, the corneas were scraped with a scalpel. Next, the central part of the cornea was cut out and minced with a scalpel. Corneal pieces were then digested with 2mg/mL collagenase (Sigma-Aldrich, # C0130) diluted in DMEM/F-12 + GlutaMAX™ medium (Thermo Fisher Scientific, Waltham, MA, USA, # 31330-095) containing 2% foetal bovine serum (FBS; Thermo Fisher Scientific, # 10082-147) and 1% penicillin-streptomycin (Thermo Fisher Scientific, # 15140-122) (DMEM/F-12 2% FBS) overnight at 37˚C. Samples were centrifuged at 1500 rpm for 5 min. In order to keep the right phenotype and function of primary keratocytes (quiescent keratocytes), the pellet was resuspended in DMEM/F-12 2% FBS, as described in our previous publication, 36 in which we demonstrated, through immunocytochemistry and Western blot analyses, that keratocytes grown in DMEM/F-12 2% FBS express specific keratocyte markers, such as CD34, keratocan, lumican and ALDH. In order to differentiate primary keratocytes into fibroblasts, for further use in the in vitro fibrosis model, the pellet was resuspended in DMEM/F12 10% FBS.
Cells were cultured at 37˚C with 5% CO 2 until they reached confluency, with medium being changed every second or third day. 0.05% trypsin-EDTA (Thermo Fisher, # 15400-054) was used to detach the cells. Cells were split into a 1:2 ratio for propagation. Central keratocytes and fibroblasts in passage 4 were used throughout this study. DMEM/F-12 2% FBS or DMEM/F-12 10% FBS was used to propagate the cell cultures. DMEM/F-12 0.1% FBS was used to assess the role of ACh on production of ECM by quiescent keratocytes. DMEM/F-12 10% FBS was used for the in vitro corneal fibrosis model. The corneas were assessed individually; that is, keratocytes isolated from different corneas were not pulled together.

| Cell viability assay
The effect of ACh on the viability of corneal fibroblasts during the onset of fibrosis was measured using MTS assay (CellTiter 96® Aqueous One Solution Cell Proliferation Assay; Promega #G3580) according to the manufacturer's instructions. Briefly, corneal fibroblasts were seeded at a density of 2 × 10 3 /well in a 96-well plate and incubated overnight. Next day, fibrosis was induced, and desired wells were concurrently treated with 10 −7 M or 10 −8 M ACh. Data were collected at times 0, 2 days and 4 days after treatment.

| BrdU incorporation ELISA
The effect of ACh on corneal fibroblast proliferation was performed by measurement of BrdU incorporation in newly synthesized cellular DNA according to the manufacturer's instructions (R&D, #11647229001). Briefly, corneal fibroblasts were seeded at a density of 2 × 10 3 /well in a 96-well plate and incubated overnight.
Next day, fibrosis was induced, and desired wells were concurrently treated with treated with 10 −7 M or 10 −8 M ACh. Data were collected at times 0, 2 days and 4 days after treatment.

| RT-qPCR
Primary keratocytes were seeded into 6-well plates at a density of Samples were run in duplicates in ViiA™ 7 Real-Time PCR System (Thermo Fisher). 18S and β-actin probes served as endogenous controls (Thermo Fisher; # 4333760F and # 4352935E, respectively).
For analysis, each time point was compared to time 0 (set to fold 1). Analysis was performed with ViiA™ 7 Software (Thermo Fisher).

| Western blot
Primary keratocytes were seeded into 6-well plates at a density of 0.25 for one hour at room temperature. Images were taken by Odyssey® Fc imaging system (LI-COR).

| Statistical analysis
All experiments were performed in triplicates. Data are presented as mean ± SD. Statistical analysis was performed with one-way ANOVA, with Tukey's post hoc test, or unpaired t test. Differences were considered statistically significant at a P-value of < .05. All experiments were performed at least three times, meaning that at least three separate experiments were performed with cells isolated from different patients (biological replicates).  Figure 1B). The same effect of ACh was observed on α-SMA protein expression; that is, both 10 −8 M and 10 −7 M ACh reduced its expression 8 days after treatment ( Figure 1C). We performed a scratch assay in order to assess whether ACh has an effect on keratocyte migration. The results showed that 10 −8 M ACh decreased keratocyte migration by 40%. However, 10 −7 M ACh had no effect ( Figure 1D). Additionally, intracellular levels of pro-collagen I, collagen III, collagen V, lumican and FN were assessed. Except for 10 −8 M ACh, which decreased pro-collagen I levels at day 8, and both 10 −8 M and 10 −7 M ACh, which reduced intracellular FN at day 4, no significant differences were found after ACh treatments ( Figure S1). showed that expression of all the genes tested was increased after treatment with VitC and TGF-β1 at days 2 and 4 ( Figure 2A). Next, we checked for secretion of pro-collagen I, collagen III, collagen V, lumican and FN. Again, induction of fibrosis with VitC and TGF-β1 resulted in increased secretion of all markers mentioned at days 2 and 4, except for collagen III secretion at day 2, for which we found no difference between treated and untreated cells ( Figure 2B). Expression of α-SMA was assessed by Western blot and flow cytometry and showed that cells treated with VitC and TGF-β1 expressed more α-SMA protein than untreated cells both at day 2 and day 4 ( Figure 2C and Figure S2F, respectively). Intracellular levels of pro-collagen I, collagen III, collagen V, lumican and FN were increased after fibrosis induction ( Figure S2).

| In vitro human corneal fibrosis model
Lastly, gel contraction assay was used to assess the contractile abilities of newly formed myofibroblasts. Induction of fibrosis with VitC and TGF-β1 resulted in significantly increased contractile abilities of the cells from day 1 to day 4 ( Figure 2D). We concluded that the in vitro human corneal fibrosis model is appropriate for our further studies on the ACh effect on the onset of fibrosis.

| Acetylcholine reduces gene expression and production of ECM components during the process of fibrosis
First, in order to assure that the concentrations of ACh used in the fibrosis study are not toxic and that the possible effect of ACh on the onset of fibrosis is not caused by cell death or lack of proliferation of corneal fibroblasts, we performed cell viability assays and BrdU incorporation ELISA. In accordance with our previous findings on ACh enhancing proliferation in primary keratocytes, 37 we also observed similar effect on corneal fibroblasts during the onset of fibrosis. Both cell viability and proliferation were enhanced in ACh-treated cells ( Figure S3). Next, we wanted to assess whether ACh affects transition of fibroblasts into myofibroblasts during the fibrosis process. First, we determined the role of ACh in gene expression of ECM components collagen I, collagen III, collagen V and lumican by qRT-PCR. The results showed that expression of COL1A1, COL3A1 and COL5A1 decreased after treatment with both 10 −8 M and 10 −7 M ACh at days 2 and 4 during the fibrosis process. Gene expression of LUM was unaffected by ACh at day 2, but it decreased at day 4 ( Figure 3A). Next, we assessed secretion of procollagen I, collagen III, collagen V and lumican by ELISA. We observed that secretion of pro-collagen I was reduced after treatment with both  Figure S4).

| Acetylcholine decreases gene expression and production of fibrotic markers during the process of fibrosis
Next step in assessing the role of ACh in transitioning of fibroblasts into myofibroblasts during the fibrosis process was to determine Acetylcholine had no effect 4 days after ACh treatment ( Figure 4C).
Additionally, scratch assay was performed to assess the effect of ACh on fibroblast migration during the onset of fibrosis. The results showed that neither 10 −8 M ACh nor 10 −7 M ACh had an effect on fibroblast migration ( Figure 4D).

| Acetylcholine inhibits corneal fibroblasts from developing contractile abilities
As our results suggest that ACh decreases transformation of corneal fibroblasts into myofibroblasts during the fibrosis process, we opted for determining whether ACh would also decrease contractile abilities of the transitioning cells. To achieve that, we used a collagen I contraction assay. We tested two concentrations of ACh-only treatment ( Figure 5C).

| Acetylcholine decreases gene expression and production of fibrotic markers in persistent fibrosis
We were interested in whether ACh exhibits the same anti-fibrotic effect in a setting where the fibrosis has already occurred. Therefore,  collagen I as it has been reported that keratocytes in cell culture are not able to process all pro-collagen I to mature collagen I, and as a result, pro-collagen I accumulates in the cell medium with only a small portion being processed to mature collagen I. 44 Moreover, ACh had no effect on collagen III secretion and we could not detect secreted collagen V. Interestingly, we have shown that ACh stimulates keratocytes to proliferate 37 ; therefore, we hypothesized that ACh would possibly increase expression of ECM and fibrotic markers.
Surprisingly, our results showed the contrary, and the reason for that should be studied.
The role of ACh in fibrosis has been investigated in airway diseases and in liver fibrosis, for which it has been shown that ACh is involved in ECM remodelling and proliferation of myofibroblasts, [39][40][41] but no data exist on its possible role in corneal fibrosis. We have adapted an in vitro human corneal fibrosis model from Karamichos et al 42 in order to study the role of ACh in the onset of fibrosis and its effect on the transition of fibroblasts to myofibroblasts. In this model, vitamin C and TGF-β1 are used to induce fibrotic process in corneal fibroblasts. Vitamin C has been shown to induce synthesis and secretion of ECM components, 45 especially collagen I, without altering non-collagen protein synthesis such as fibronectin. 46 Transforming growth factor-β1 stimulates overproduction and deposition of ECM. [8][9][10] Using this model, we were able to induce overproduction of ECM components (collagen I, collagen III, collagen V and lumican) and expression of the fibrotic markers FN and α-SMA in newly formed myofibroblasts, which presence was confirmed by gel contraction assay.
Again, our results showed that ACh had an inhibitory effect on the formation of fibrosis and myofibroblasts in our model. It down-regulated gene expression of collagen I, collagen III, collagen V and lumican. Secretion of pro-collagen I and lumican was decreased; however, secretion of collagen III was unaffected by ACh. Interestingly, secretion of collagen V was enhanced by ACh.
However, the results showed that the intracellular content of collagen V was decreased by ACh. Perhaps, the ACh-treated cells were Values are means ± SD. n.s. (not significant); *P < .05; **P < .01; ***P < .001; ****P < .0001 able to secrete collagen V more rapidly, but the total content of collagen V (intracellular + secreted) was unaffected by ACh treatment, as was the case for collagen III, for which ACh had no effect on either its secretion or its intracellular levels. Moreover, ACh decreased gene expression and secretion of FN, which might explain the lower levels of secreted collagen I, as FN is responsible for its deposition during fibrosis. 47 We have found no apparent difference in action between the two concentrations of ACh used. Our results also suggest that the anti-fibrotic effect of ACh is mediated by activation of mAChRs. In our previous study, we found that ACh treatment in keratocytes activates mAChRs rather than nAChRs, 37 in order to enhance keratocyte proliferation. Perhaps, the activation of mAChRs, rather than nAChRs, in keratocytes and corneal myofibroblasts leads to an anti-fibrotic response, rather than a pro-fibrotic one, as it has been reported for airways that activation of α7 nAChR is involved in the progression of lung fibrosis. 39,40 Studies have shown contradicting results regarding the role of the mAChR antagonist atropine in fibrosis. Atropine has been found to have an anti-fibrotic effect in hepatocytes isolated from rat fibrotic liver 48 ; however, others have shown that atropine has pro-fibrotic properties in rat cardiac fibroblasts. 49 Our results suggest that atropine alone does not affect the fibrotic process in corneal fibroblasts. However, it is possible that it affects collagen secretion.
Finally, we showed that ACh hinders transition of corneal fibroblasts into myofibroblasts, as demonstrated by down-regulation of α-SMA and subsequent inhibition of development of contractile activity by the fibroblasts. Again, this effect was shown to be mediated by activation of mAChRs, since the usage of atropine increased the contractile activity of the cells. The effect of ACh was dose-dependent with higher concentration inhibiting the contraction more.
Importantly, we showed that ACh (only the lower concentration) decreased keratocyte migration, which further supports our hypothesis and shows that ACh has an effect under more physiological settings, as mechanical wounds or chemical burns are one of the most common reasons for corneal injury. We think that during injury, ACh affects the keratocytes in a way that they remain quiescent or that they are activated in a slower manner. The reason for ACh not having an effect on fibroblast migration during the onset of fibrosis is perhaps because of the experimental settings, that is very high amount of FBS, which under physiological conditions would be omitted.
The small, but significant, decrease in fibrotic markers by ACh observed in this study could have potential clinical application. The physiological conditions differ greatly from the in vitro settings, and at least up to this date, it is hard to reproduce the physiological conditions. Issues such as substrate stiffness will affect the in vitro experiments, and isolation of the cells from their natural environment will, to a degree, change them too. 50,51 We expected to see a significantly bigger decrease in α-SMA expression, as contraction of the collagen gels, in which the corneal fibroblasts are embedded, was decreased greatly by the ACh-treated cells. Perhaps, α-SMA removal from the stress fibres is somehow deficient, or α-SMA is not properly degraded. One study showed that when inducing apoptosis in fibroblasts, α-SMA was degraded by caspase-3. 52 We have previously shown that ACh has an anti-apoptotic effect on keratocytes. 38 Therefore, we could speculate that a similar mechanism inhibits α-SMA degradation in this study. Additionally, as ACh increases proliferation of corneal fibroblasts during the onset of fibrosis, the small changes observed could be a result of that. Perhaps treating the cells with ACh after arresting the cell cycle could result in a bigger and clearer change.
Taken together, our results suggest that ACh displays anti-fibrotic characteristics in an in vitro human corneal fibrosis model. ACh impedes overproduction of ECM components and hampers expression of fibrotic markers, and this effect is driven by activation of mAChRs.
Therefore, ACh not only might play a regulating role during the initial stages of corneal fibrosis, but also might decrease an already-existing fibrosis. Hence, perhaps it contributes to reduced scarring of the cornea. Our findings are promising, as it has been reported by Uberti et al 53,54 that kinetically energized ultra-low doses of ACh show remarkably great wound healing properties both in vitro in human keratinocytes and in vivo in mice. Even though our results cannot be directly compared with Uberti's findings, throughout our studies on ACh in corneal wound healing, we have observed that the lower dose of ACh exerts stronger wound healing and anti-fibrotic effects.
Moreover, it seems that the kinetically energized ultra-low-dose ACh is safe for topical applications to skin in pre-clinical studies, which would be very promising and beneficial for treating corneal wounds. Additionally, it might be interesting to study whether application of anti-cholinergic drugs such as tropicamide, atropine or cyclopentolate, which are commonly administered for the purpose of ocular examination in order to dilate the pupil, 55 or application of cholinergic drugs such a pilocarpine, which is used to constrict the pupil in the treatment of angle closure glaucoma, 56

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.