Effects of Trehalose on VEGF-Stimulated Angiogenesis and Myofibroblast Proliferation: Implications for Glaucoma Filtration Surgery

Abstract

Purpose.: To investigate whether trehalose inhibits VEGF-stimulated or inflammatory angiogenesis and the proliferation of myofibroblasts.

Methods.: Normal human dermal fibroblasts and human umbilical vein endothelial cells (HUVECs)were cocultured in trehalose-containing medium (2.5/5.0/7.5/10.0%) with or without VEGF (10 ng/mL). After 11 days, the area, length, joint, and path of neovascularization were evaluated. The effect of topical trehalose on corneal neovascularization was examined in vivo by treating Balb/c mice with alkali burn-induced corneal neovascularization. After 14 days of trehalose treatment, corneal vessels were visualized in flatmounts. The expressions of VEGFR2, phospho-VEGFR2, and vimentin were observed. Then a separate coculture model of the myofibroblasts and HUVECs was used to observe the morphologic changes of the myofibroblasts by trehalose. Furthermore, myofibroblasts were cultured with trehalose to examine the cytokeratin and E-cadherin expressions.

Results.: In the in vitro models, there was a significant trehalose dose-dependent inhibition of neovascularization. In the in vivo alkali burn models, corneal neovascularization was significantly inhibited by treatments using ≥2.5% trehalose eyedrops. The expressions of VEGFR2, phospho-VEGFR2, and vimentin were downregulated by trehalose. When trehalose was added to the medium, the myofibroblasts were transformed into epithelial cell-like cells. The transformed myofibroblasts expressed cytokeratin but not E-cadherin.

Conclusions.: Trehalose prevents angiogenesis by partially downregulating VEGFR2 expression. In addition, trehalose inhibits the proliferation of myofibroblasts partially by inducing mesenchymal-epithelial transition. These findings suggest that trehalose has potential for use as a new agent that can control angiogenesis and fibrosis and potential for use in glaucoma surgery.

Several overlapping reactions are responsible for wound healing in adult skin, conjunctival wounds after glaucoma filtration surgery (GFS), and ocular surface neovascularization. The wound healing processes that occur in both adult skin and after GFS consist of an acute inflammatory phase in which there is infiltration of inflammatory cells and the upregulation of several cytokines, a proliferative phase in which there is migration of fibroblasts and vascular endothelial cells from the neighboring tissues and the transformation of fibroblasts into myofibroblasts, and a remodeling phase that eventually leads to the formation of fibrous scar. Recently, the effects of cytokines on the pathogenesis of fibrous scar formation have been increasingly investigated. Several studies have examined transforming growth factor β (TGF-β), ^1,2 connective tissue growth factor, ³ platelet-derived growth factor, ⁴ bone morphogenic proteins, ⁵ and vascular endothelial growth factor (VEGF). ^{6 –9} Usually, excessive fibrous scar formation in the surgical wound area is a frequent cause of failure of GFS. Given that angiogenesis plays an important role in the formation of granulation tissue and subsequent fibrous proliferation, the inhibition of angiogenesis is a potential therapeutic way to control excessive fibrous scar formation after GFS in addition to the direct inhibition of fibrous proliferation. This idea has been supported by the report that subconjunctival injection of anti-VEGF antibodies administered immediately after trabeculectomy revealed significant reduction of bleb failure after surgery. ^{10 –13} Furthermore, it is likely that inhibition of VEGF activity may control scarring by inhibiting both fibroblast proliferation and fibroblast contractile ability and by inducing fibroblast death (Qin Q, et al. IOVS. 2008;49:ARVO E-Abstract 4541). These findings suggest that there is an interaction between fibrous proliferation and angiogenesis in wound tissue. In addition, VEGF not only induces vascular proliferation, it also acts as a mediator in the signal pathway that leads to fibroblast proliferation. ^14,15 Because VEGF is upregulated in the aqueous humor in patients with glaucoma, postoperative fibrous scar formation after GFS could potentially be accelerated by VEGF. ¹² A clinical trial designed to examine the efficacy of topically applied ranibizumab eyedrops on fibrous scar formation after GFS is under way. ¹⁶  

In a previous study, ¹⁷ we demonstrated that trehalose inhibited the proliferation of fibroblasts by inducing apoptosis. In rabbit GFS models, we also observed that trehalose inhibited conjunctival neovascularization in bleb areas after GFS. ¹⁷ If trehalose has an inhibitory effect on VEGF-induced vascular proliferation, it can be safely used as an adjunctive agent to maintain the filtration bleb in combination with the mitomycin C or 5-fluorouracil, which are normally used during and after GFS. Based on previous findings that have been reported, ¹⁷ we designed a new study to further investigate the effects of trehalose on the proliferation of vascular endothelial cells. This study used an in vitro cell coculture system and an experimental in vivo corneal neovascularization model that was designed to examine possible interactions between myofibroblasts and vascular endothelial cells. 

Materials and Methods

Culture Materials and Animals

We purchased angiogenesis kits from Kurabo Industries (Osaka, Japan). This kit consisted of a cocultured 24-well plate model that admixed and seeded normal human dermal fibroblasts (NHDFs) and human umbilical vein endothelial cells (HUVECs). Given that neither fibroblasts nor vascular endothelial cells derived from normal human Tenon space are commercially available, we simulated these cells in the subconjunctival connective tissue using NHDFs and HUVECs, as previously reported. ¹⁷ Cell culture dishes, plates, centrifuge tubes, and other plastic ware were purchased from Nunc (Thermo Fisher Scientific, Roskilde, Denmark). 

All animal experimental procedures were designed ethically to conform to both the ARVO Statement for the Use of Animals in Ophthalmic Vision Research and the guidelines of our own institution. Balb/c mice were purchased from Japan Clea (Tokyo, Japan). For anesthesia, mice were first administered diethyl ether by inhalation, followed by intramuscularly injected ketamine hydrochloride (80–125 mg/kg) and xylazine (9–12 mg/kg) once the animals were unconscious. All surgical manipulations were performed under additional topical anesthesia that used 0.4% oxybuprocaine hydrochloride eyedrops. Right eyes were used in all the animal experiments. 

Angiogenesis in an In Vitro Model

To examine the effect of trehalose on angiogenesis, HUVECs were cocultured with NHDFs in the angiogenesis kit with or without VEGF and trehalose (Nacalai Tesque, Kyoto, Japan), respectively. The mixture of NHDFs and HUVECs was incubated for up to 11 days at 37°C in a 5% CO₂ incubator. After 1, 4, 7, and 9 days of culture, the media were replaced by freshly prepared trehalose-containing media (2.5%/5.0%/7.5%/10.0%) with or without 10 ng/mL VEGF (Kurabo). Suramin (50 μM; Kurabo) was used as the inhibitor of VEGF. The media conditions we used were divided into four groups as follows: media without VEGF and suramin as a negative control, media with VEGF and without suramin as a positive control, media with both VEGF and suramin, and media with both VEGF and 2.5% to 10.0% trehalose, respectively (n = 4 in all groups). The trehalose-containing culture media contained 2% fetal bovine serum (Kurabo). 

After 11 days of coculture, cells were fixed at room temperature in 70% ethanol that had been stored at −20°C. Subsequently, cells were blocked with 1% bovine serum albumin (BSA) in phosphate-buffered saline for 30 minutes at room temperature and then were incubated with mouse anti–human platelet-endothelial cell adhesion molecule-1 (PECAM-1, [CD31]) antibody, 1:40; Kurabo) as the primary antibody for 1 hour at 37°C. The cells were then further incubated with donkey anti–sheep immunoglobulin G (IgG) conjugated by alkaline phosphatase (1:500) (Kurabo) as the secondary antibody for 1 hour at 37°C. The endothelial cells were visualized by nitro-blue tetrazolium (NBT)/5-bromo-4-chloro-3′-indolyl phosphate (BCIP). 

To quantify the extent of proliferation and differentiation of the endothelial cells, we defined four parameters of the vascular network in each well, as follows: area, which was defined as the total area of the tube formation stained by NBT and BCIP (expressed by pixels); length, which was the total length of the DAB-stained tube network; joint, which was the total number of bifurcations in the vascular network; and path, which was defined as the total number of vascular branches from the bifurcation. Area and length were measured by using ImageJ software (developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html), and statistical analyses were performed by angiogenesis quantitative analysis software (Kurabo). 

In Vivo Corneal Neovascularization Model

We created alkali burn–induced corneal neovascularization using Balb/c mice. After placing a filter paper (1.5 mm in diameter) presoaked in 1 N NaOH on the right cornea of each animal for 60 seconds, we randomly divided the mice into four groups. Three groups received trehalose eyedrops (2.5%, 5.0%, 7.5%) and the fourth group received 0.9% saline eyedrops in triplicate. Topical application of the eyedrops continued four times a day for 14 days. To quantify the corneal neovascularization in corneal flatmounts, after 14 days, intraventricular (left cardiac ventricle) injections of 0.7 to 1.0 mL fluorescein-isothiocyanate-dextran (Sigma-Aldrich, St. Louis, MO) was injected in the mice, followed by enucleation and observation by fluorescein microscopy. Corneal flatmounts were photographed, and the areas of corneal neovascularization were quantified. Briefly, flat-mounted corneal photographs were divided into 10 equally wide fan shapes (36° of the top angle). Neovascularization areas for each of the fan shapes were measured using ImageJ software and were expressed by pixels. Differences in the neovascularization areas for all the groups were then statistically compared. 

We also used Western blot analysis to examine the effects of trehalose on VEGF receptor 2 (VEGFR2), phospho-VEGFR2, and vimentin protein expression in the alkali burn–induced cornea. For this assay, 10 eyeballs of the Balb/c mice in each of the trehalose eyedrop groups and the control group were enucleated, with the corneas then removed from the eyes by dissection and homogenized by a sonicator in lysis buffer (1× Roche complete mini protease inhibitor cocktail; Roche Applied Science, Indianapolis, IN). The homogenates were subsequently used for Western blot analysis. 

NHDF/HUVEC Membrane–Separated Coculture Model

To evaluate the effects of trehalose on myofibroblasts cocultured with HUVECs, we first cultured NHDFs in 106S medium containing 2% fetal bovine serum for more than 2 months. Immunohistochemistry and Western blot analysis showed that most of the fibroblasts were transformed to α-SMA–positive myofibroblasts (Supplementary Fig. S1, http://www.iovs.org/lookup/suppl/doi:10.1167/iovs.11-7478/-/DCSupplemental). These myofibroblasts were then placed into each well of a 24-well plate (5 × 10³ cells/well), with a basket-type culture well unit (0.45-μm diameter pore; Intercell; Kurabo) put into each well of the culture plates (Supplementary Fig. S2, http://www.iovs.org/lookup/suppl/doi:10.1167/iovs.11-7478/-/DCSupplemental). Then HUVECs (2 × 10⁴ cells/well) were inoculated into the culture well unit (Intercell; Kurabo) and separately cocultured with the myofibroblasts. The micropore membrane that was located at the bottom of the culture well unit prevented direct mixture of the myofibroblasts and HUVECs and allowed only soluble substances to penetrate. We used 200S medium (Kurabo) for HUVECs in this coculture model and observed the morphology of the myofibroblasts in media with or without trehalose. 

Effects of Trehalose on Myofibroblasts

NHDFs and HUVECs were cultured separately with the HUVEC medium (200S; Kurabo), then extracted and concentrated 20 times using concentrators (Vivapore; Sartorius Stedim Laboratory Ltd., Louth, UK). NHDFs were cultured for approximately 2 months until the fibroblasts transdifferentiated into myofibroblasts. The cells were then inoculated into each well plate (5 × 10³ cells/well) of a four-chamber slide (Laboratory-Tek; Nalge Nuc International, Naperville, IL) in 10 mL of 200S culture media, with 1.0 mL of the 20× concentrated medium derived from the cultured HUVECs added to the medium of the myofibroblasts with or without trehalose. The cells were further cultured for 1 week and then harvested for use with the immunocytochemistry and Western blotting procedures, which were used to detect cytokeratin, E-cadherin, and vimentin protein expressions. For Western blot analysis, cells were dissolved in lysis buffer (1× complete mini protease inhibitor cocktail; Roche Applied Science). The respective protocols are described in the sections that follow. 

Immunocytochemistry

The immunocytochemistry procedure for the cytokeratin, E-cadherin, and vimentin in the cultured myofibroblasts used cells that were fixed in mixed 100% ethanol and 100% acetone at the same volume. These were then blocked with 0.5% skim milk in PBS containing 0.05% of Tween 20 (0.05% TW-PBS) for 1 hour at room temperature. The slides were treated with mouse anti–human cytokeratin antibody (1:200) (Progen Biotechnic, Heidelberg, Germany), rabbit anti–human E-cadherin antibody (1:200) (Santa Cruz Biotechnology, Santa Cruz, CA), or rabbit anti–human vimentin antibody (1:200; Epitomics, Burlingame, CA) as the primary antibody overnight at 4°C. The slides were treated with Alexa Fluor 568 anti–mouse IgG conjugate (1:200; Molecular Probes, Eugene, OR) and swine anti–rabbit FITC conjugate (1:200; DAKO, Glostrup, Denmark) overnight at 4°C. Nuclei were counterstained with 4′,6′-diamino-2-phenylindole (DAPI). Tissue sections were examined under a fluorescence microscope. 

Western Blot Analysis

Western blot analysis was used to evaluate protein expression for VEGFR2, phospho-VEGFR2, and vimentin in the alkali-burned mice corneas and for E-cadherin and cytokeratin in the cultured myofibroblasts. Briefly, the same amount of protein (20 or 25 μg) from the tissue homogenates or cell lysates was loaded in each lane of a 12% acrylamide gel and then electrophoresed and electrotransferred onto polyvinylidene difluoride membranes (Bio-Rad, Hercules, CA) at 18 V overnight. After nonspecific binding was blocked by a blocking buffer that was either 0.5% (wt/vol) skim milk or 5% BSA containing 1% normal goat serum in 0.05% TW-PBS at room temperature for 2 hours. The membrane was then probed with primary antibody against mouse VEGFR2, mouse phospho-VEGFR2 (Cell Signaling Technology, Danvers, MA), mouse vimentin (Epitomics), human E-cadherin (Santa Cruz Biotechnology), and mouse pan-cytokeratin (Progen Biotechnik, Heidelberg, Germany) diluted at 1:1000 in blocking buffer at 4°C overnight. After washing three times with 0.05% TW-PBS, the membrane was incubated in blocking buffer containing horseradish peroxidase (HRP)-labeled goat-anti–rabbit or mouse IgG (DAKO) diluted at 1:5000 at 4°C overnight. Specific antigen-antibody binding was visualized using an enhanced chemiluminescence (ECL) kit (GE Healthcare, Little Chalfont, UK). After detection of the band for each protein, a Western blot recycling kit (Re-Blot; Chemicon International, Temecula, CA) was used to remove the anti–antibody. The membrane was then blocked by 0.5% skim milk and probed with primary antibody against rabbit anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Abcam, Cambridge, UK) as an internal protein control diluted to 1:5000 in blocking buffer at 4°C overnight. After washing 3 times with TW-PBS, the membrane was incubated with HRP-labeled goat anti–rabbit IgG (DAKO) diluted at 1:5000 at 4°C overnight. Specific antigen-antibody binding GAPDH was visualized using an ECL kit. 

Statistical Analysis

One-way ANOVA and an unpaired Student's t-test were used to evaluate the effect of trehalose on angiogenesis and fibrous scar formation by using the SPSS statistical software (version 17; SPSS Inc., Chicago, IL). The significance level was set at P < 0.05. 

Results

Inhibition of Angiogenesis by Trehalose in an In Vitro Model

When NHDFs and HUVECs were cocultured in the presence of VEGF and CD31⁺ vascular endothelial cells, significantly more proliferation was observed compared with the control medium (Fig. 1). Results also showed that VEGF stimulated vascular tube formation within all the angiogenesis categories including area, length, joint, and path (Fig. 2). For the negative control, suramin significantly inhibited the proliferation of CD31⁺ vascular endothelial cells, even in the presence of VEGF (P < 0.01 for all parameters). When NHDFs and HUVECs were cocultured in the presence of VEGF and trehalose (2.5%/5.0%/7.5%/10.0%), there was a significant trehalose dose-dependent inhibition of the proliferation of CD31⁺ vascular endothelial cells (Figs. 1, 2). In particular, compared with suramin, 7.5% and 10.0% trehalose significantly inhibited the proliferation of vascular endothelial cells in two categories, joint and path (P < 0.01). The time courses of the effect of trehalose in area and length are shown in Figure 3, indicating that the inhibitory effect of trehalose is correlated to the duration of culture up to 11 days. In addition, when the background fibroblasts were in the presence of trehalose (2.5%/5.0%/7.5%, respectively; Fig. 1), they morphologically changed into phenotypes such as epithelial cells. 

Figure 1.

 

Morphologic changes of HUVECs in the presence of VEGF (10 ng/mL) and trehalose. Culture conditions of each group in descending order are VEGF (−), suramin (−), trehalose (−); VEGF (+), suramin (−), trehalose (−); VEGF (+), suramin (50 μM), trehalose (−); VEGF (+), suramin (−), trehalose (2.5%); VEGF (+), suramin (−), trehalose (5.0%); VEGF (+), suramin (−), trehalose (7.5%); and VEGF (+), suramin (−), trehalose (10.0%). Tube formation by endothelial cells is promoted in the presence of VEGF but inhibited by suramin (VEGF inhibitor) and trehalose. Original magnifications: (left) 40×; (center, right) 100×.

Figure 2.

 

(a–d) Statistical analyses performed to evaluate the area (a), length (b), path (c), and joint (d). Significant inhibition of the VEGF-induced angiogenesis is seen for area, length, joint, and path in the presence of 2.5% (P < 0.01), 5.0% (P < 0.01), 7.5% (P < 0.01), and 10.0% (P < 0.01) trehalose concentrations (t-test). **P < 0.01. Bars indicate standard deviations.

Figure 3.

 

Relationships between concentration of trehalose and incubation time course in area (a) and length (b). Incubation periods: light gray, 3 days; dark gray, 6 days; black, 11 days.

Inhibition of Corneal Angiogenesis by Trehalose in an In Vivo Model

We created alkali burn–induced corneal neovascularization in Balb/c mice. The advantage of this experimental model is that it is possible to perform direct microscopic observation of the neovascular networks at 14 days. When mice were treated postoperatively with trehalose eyedrops, there was a significant dose-dependent inhibition of the elongation of neovascular network (Figs. 4, 5). There were significant differences noted for the areas of corneal neovascularization between the control and 2.5% to 7.5% trehalose groups (one-way ANOVA; P < 0.001). Individual analyses revealed significantly fewer areas of neovascularization for the 2.5% (P = 0.023), 5.0% (P = 0.001), and 7.5% (P < 0.001) trehalose groups compared with the controls (unpaired t-test; Fig. 5). As seen in Figure 6, instillation of trehalose for 14 days after surgery resulted in a downregulation of the VEGFR2, phospho-VEGFR2, and vimentin protein expressions in the alkali-burned corneas. 

Figure 4.

 

Fluorescein microscopic views of flatmounted corneal alkali burn–induced neovascularization and the effects of a topical instillation of trehalose eyedrops (2.5%–7.5%, respectively). PBS as control.

Figure 5.

 

Quantitative analyses of corneal neovascularization areas. Measurements of the neovascular area in each of the 36° top-angle fan shapes by ImageJ, with expression in pixels (n = 30, each group). The concentrations of 2.5% (P = 0.023, t-test), 5.0% (P = 0.001, t-test), and 7.5% (P < 0.001, t-test) trehalose significantly inhibited corneal neovascularization when compared with control (P = 0.001, one-way ANOVA). *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 6.

 

Western blot analyses of VEGFR-2, phospho-VEGFR-2, and vimentin expression in the alkali-burned mice corneas after exposure to different postoperative instillation concentrations of trehalose. Lane 1, PBS (control); lane 2, 2.5% trehalose eyedrops; lane 3, 5.0% trehalose eyedrops; lane 4, 7.5% trehalose eyedrops; and lane 5, 10.0% trehalose eyedrops. For the different trehalose concentrations, there was a dose-dependent reduction of the expressions of VEGFR-2, phospho-VEGFR2, and vimentin.

Cytokeratin Expression in Myofibroblasts

When NHDFs and HUVECs were cocultured in the coexistence of VEGF and trehalose, not only was angiogenesis inhibited but the morphologic changes of the fibroblasts into an epithelial cell-like form were observed (Figs. 2, 7a). This notable phenomenon was also specifically observed when myofibroblasts were separately cultured with HUVECs in a culture well unit with 200S media that contained both VEGF and trehalose (Fig. 7b). To determine whether this morphologic transformation of the myofibroblasts was a mesenchymal-epithelial transition (MET), we used both immunocytochemistry and Western blot analysis to test for cytokeratin and E-cadherin protein expressions. Western blot analysis demonstrated that only cytokeratins (CK16, CK17, and CK18) were expressed in the myofibroblasts (Fig. 8), whereas E-cadherin was not detected (data not shown). Cytokeratins were expressed even in culture media that contained 5% trehalose without VEGF or supernatant from HUVECs (Fig. 8), suggesting that trehalose promotes transdifferentiation of the myofibroblasts into an epithelial cell-like morphology. This pattern was confirmed by immunocytochemistry, for which the myofibroblasts expressed both vimentin and cytokeratin but not E-cadherin in the cells (Fig. 9). 

Figure 7.

 

(a) Morphologic changes of the fibroblasts in the NHDF and HUVEC coculture models (magnified from Fig. 2). (left) Medium that contained VEGF (10 ng/mL) without trehalose. (right) Fibroblasts and HUVECs in the medium containing 7.5% trehalose and VEGF (10 ng/mL). Epithelial cell-like morphologic change of fibroblasts are seen. Scale bar, 50 μm. (b) Morphologic changes of the myofibroblasts when using the culture well unit system to coculture the cells separately with HUVECs (Fig. 2). Changes of the myofibroblasts to the epithelial cell-like form are also seen when cultured with trehalose. Scale bar, 50 μm.

Figure 8.

 

Western blot analyses for cytokeratins in the myofibroblasts cultured with trehalose. Lane 1, trehalose-free medium; lane 2, 5.0% trehalose, VEGF (10 ng/mL), and supernatant from HUVECs; lane 3, 5.0% trehalose and supernatant from HUVECs; and lane 4, 5.0% trehalose medium only.

Figure 9.

 

Evaluation of cytokeratin, vimentin, and E-cadherin protein expressions by immunocytochemistry. Images show the myofibroblasts cultured in the media containing 5.0% trehalose and supernatant derived from HUVECs. Although cytokeratin is expressed in the myofibroblasts, results are negative for E-cadherin, which confirms the Western blot results. Scale bars, 60 μm.

Discussion

Trehalose is a disaccharide that is an isomer of sucrose and that has a molecular weight of 342.29 Da. As trehalose is regarded as safe for use in humans; it now widely used an ingredient in foods, cosmetics, and medicines. Recently, it also has been suggested that trehalose may be a safe topical agent for effectively treating dry eye ¹⁸ and be applicable as a preservative agent to keep amniotic membranes viable during freezing conditions. ¹⁹ In our previous study, ¹⁷ we reported that trehalose also inhibited the proliferation of fibroblasts by inducing them into apoptosis. These studies all suggest that trehalose eyedrops might be able to prevent abnormal fibroblast proliferation by safely preserving ocular surface epithelial cells during GFS and pterygium resection. Although antimetabolic agents such as 5-fluorouracil and mitomycin C are used to control postoperative subconjunctival fibrous scar formation after GFS, effects of these antimetabolites tend to be irreversible and can cause serious blinding complications such as avascular filtration bleb, conjunctival buttonhole formation, and subsequent infection of the filtering bleb that leads to endophthalmitis and scleral necrosis during the later stages. ^20,21 Therefore, to ensure there is an effective and safe post-GFS filtering bleb that will reduce blindness due to glaucoma, the development of further adjunctive therapy or agents that can replace the use of antimetabolites is required. The increasing number of investigations on the wound healing process has led to the discovery of important roles that cytokines play in postoperative fibrous scar formation. TGF-β has been shown to promote the transdifferentiation of fibroblasts into myofibroblasts that ultimately produce excessive fibrous scar. However, subconjunctival injections of a specific dose of an antibody against TGF-β (CAT-1520102 Trabeculectomy Study Group) after GFS were not able to prevent the failure of trabeculectomy in a multicenter clinical trial. ²² Although this result might have been partially due to an insufficient dosage of antibody, the importance of VEGF has attracted much attention with regard to its wound modulatory properties. ²³ Because angiogenesis takes place before fibrous scar formation, VEGF may play a role in the conjunctival scar formation that occurs after GFS. Therefore, the key to opening a new therapeutic pathway for regulating fibrous scar proliferation after GFS may center on controlling the VEGF effect on angiogenesis. In line with this, the potential of anti–VEGF treatments after GFS have been increasingly investigated (Qin Q, et al. IOVS. 2008;49:ARVO E-Abstract, 4541). ^{10 –13,16}  

In the present study, we demonstrated that trehalose inhibited both VEGF-induced angiogenesis in vitro (Figs. 2, 3) and corneal inflammation in vivo (Figs. 4, 5). Although the exact molecular mechanisms remain unknown, trehalose downregulates VEGFR2 and phospho-VEGFR2 expressions in the inflammatory lesion cells (Fig. 6). Although this may partially explain the molecular mechanism of trehalose, this activity may also be an indirect result of decreased tissue vascularity. Further investigations of the intracellular molecular changes that take place after trehalose treatment will have to be conducted. 

Previous studies have demonstrated that interactions among fibroblasts, myofibroblasts, and vascular endothelial cells take place within the inflammatory lesion. ^{6 –8} In the present study, we showed that after the addition of trehalose in the culture media that contained humoral factors secreted from the vascular endothelial cells, the myofibroblast morphology changed to an epithelial cell-like form (Fig. 7). This phenomenon was observed even when the culture media did not contain either VEGF or supernatant from HUVECs (Fig. 8). MET is a biological cell development program that is the reverse of epithelial-mesenchymal transition (EMT). Both these mechanisms have been shown to play an important role in the metastatic process. In the present study, we examined the cytokeratin and E-cadherin protein expressions along with the epithelial cell markers to determine whether the transformed myofibroblasts developed MET. Results indicated that only the cytokeratins, but not E-cadherin, were expressed in the myofibroblasts (Figs. 8, 9), suggesting that incomplete MET might have occurred. However, they did not fully differentiate into epithelial cells, and they still contained mesenchymal cell properties. Vimentin, which is a mesenchymal cell marker, was expressed in the transformed myofibroblasts. Although the mechanisms remain unknown, either inhibition of the vascular endothelial cell proliferation or a direct effect of trehalose on the myofibroblasts could be responsible for the observed changes. In our previous study, we showed that trehalose inhibited the proliferation of fibroblasts by inducing apoptosis. ¹⁷ Therefore, the MET-like morphologic changes of the myofibroblasts could be a sign that trehalose inhibits the growth of cells. To definitively clarify the effects of trehalose on the growth inhibition in both the vascular endothelial cells and the myofibroblasts in the Tenon space after GFS, further investigations will have to be conducted. Given that trehalose is antigenically inert and can safely be used as eyedrops, this agent could potentially be used in the future as an adjunctive treatment to the antimetabolites and other anti–VEGF agents that are normally administered after GFS. In addition, it will be necessary to clarify in the future whether such effects on vascular endothelial cells and myofibroblasts are specific for trehalose or are commonly seen in other sugars. 

Supplementary Materials

Figure sf01, JPG - Figure sf01, JPG 

Figure sf02, JPG - Figure sf02, JPG 

Footnotes

 Supported in part by Grant-in-Aid for Scientific Research C-21592213 (MN) from the Ministry of Education, Science, Culture and Sports of the Japanese Government.

Footnotes

 Disclosure: K. Takeuchi, None; M. Nakazawa, None; Y. Ebina, None